STAR LECTURE 2020

The hidden power of mathS Presented by Matt Parker Organised by:

Featuring the Christmas Lectures from The Royal Institution, Great Britain (x+y) n n Message from Σ( n )x = k=0 Chief executive k n-k Science centre Board n-k y Assoc Prof Lim Tit Meng Chief Executive Science Centre Board

This will mark the ninth year that the Science Centre Singapore and the Agency for Science, Technology and Research (A*STAR) have come together to bring to Singapore the world-renowned Royal Institution (Ri) Christmas Lectures from the United Kingdom, drawn from a proud English tradition of Science learning since 1825.

Singapore students, youths and even adults can explore and take learning beyond the classroom through the STAR Lecture presenter. With positive reviews and feedback gathered from past attendees, we hope that the legacy will continue through this year’s lecture which will adopt a slightly n different format due to the travel restrictions in place in view of COVID-19. This year’s lecture would be recorded both in UK and Singapore before the presentations are combined. Cr Our presenter Matt Parker warns us that Written by our unwavering faith in numbers can lead to disaster when we get the sums wrong. I am sure that Matt will get more people Octave Goh He will decode and celebrate what makes excited about mathematics as he is not only our human minds so unique during this a recreational mathematician but also a Science Educator lecture and help probe the limits of Maths YouTube personality and communicator. Physical Sciences Group and its role in our world. So, be ready for a lecture which will take you on an inspirational, fun and mesmerising journey to reveal the Hidden Power of Maths! 2 STAR LECTURE 2020 STAR LECTURE 2020 3 About the y = mx + c Presenter Message from Matt Parker EXECUTIVE DIRECTOR Presenter for STAR Lecture 2020 A*STAR Graduate Academy Prof LisA Ng Executive Director, A*STAR Graduate Academy Matt Parker is a mathematician and Agency for Science, Technology and Research (A*STAR) entertainer who uses humor to communicate mathematical topics to a worldwide audience through YouTube videos, TV and radio appearances, books and newspaper articles Welcome back to another year of STAR as well as stand-up comedy. Matt originally Lecture! This is the ninth successful year of trained as a high school Mathematics this wonderful partnership between the teacher in Perth, Australia, before moving Agency for Science, Technology and to the United Kingdom. He now gives maths Research (A*STAR) and the Science Centre and physics talks for secondary schools Singapore; with the Royal Institution of students and adult audiences all across the Great Britain. Since it debuted in 2012, this UK and internationally. unique series has been immensely popular with our young audiences over the years. Ultimately, he aims to ‘get more people With this year’s show, we look forward to excited about maths’. He is popular online, unravelling the hidden power of mathematics with his YouTube channel, StandUpMaths, through an exciting variety of science having over 500,000 subscribers, and both another mathematics communicator, Matt experiments and demonstrations. Get an of his videos on the Royal Institution won the 2018 Communications Award of insider’s look at the secret algorithms, rules channel gaining well over 1 million views the Joint Policy Board for Mathematics and patterns controlling all aspects of our each. Together with Victoria (Vi) Hart, (USA). As well as BBC Radio, he is regularly world! found talking about maths on BBC News, Sky News, Channel 4, CBBC and occasionally Our speaker Matt Parker will take you on an writes for The Guardian. Matt Parker has exhilarating journey of discovery to learn also written two bestselling popular science how elements like probability determine books, ‘Things to Make and Do in the Fourth the likelihood of everyday happenings such Dimension’ and ‘Humble Pi’. as volcano eruptions. I encourage you to empower yourselves with the knowledge of mathematics and discover how it affects our future in unimaginable ways.

Enjoy the show! 4 STAR LECTURE 2020 σ STAR LECTURE 2020 5 The hidden Probability power of Maths

Impossible Unlikely Even Chance Likely Certain

0 1

1-in-6 Chance 4-in-5 Chance

Visual Representation of Probability

Probability is the branch of mathematics using numerical values to describe how likely an event or proposition is true. The numerical value varies between 0 and 1, where 0 indicates impossibility and 1 indicates certainty. We think our lives unfold thanks to a mix of luck and our own personal choices. However, we tend to overlook the Examples of probability include: layer of mathematics that governs every aspect of our world. • When a fair coin is flipped, the probability of either “heads” or “tails” The “chance” that we often speak of is often related to is 1/2. probability. The “choices” that we make are influenced by computer algorithms. • When rolling a die, the probability of getting a 3 is 1/6. Understanding these aspects helps us make better choices and sort fact from fiction. However, putting too • When drawing a card from a shuffled much trust and getting the sums wrong can lead to deck, the probability of drawing the 1/6 disaster. Aces of Spades is 1/52.

6 STAR LECTURE 2020 STAR LECTURE 2020 7 NORMAL Normal Distribution Distribution

Coin flipping is an independent event. It means that the outcome of the 2nd flip is not influenced by the outcome of the 1st flip. Regardless of the result of the 1st flip, the Graph nd probability of the 2 flip to get a “heads” is still 1/2. As we further increase the number of coin flips, the bar chart will start to take the shape of a normal distribution curve: An equation used to calculate probability of an event (A) occurring is:

Number of ways Event A can occur Probability of event A = Total number of Possible Outcomes 0.4 So, what will happen if we flip 2 coins? There is a total of 4 possible outcomes as shown in the table below: 0.3 Outcome 1st Flip 2nd Flip

1 “Heads” “Heads” 0.2 34.1% 34.1% 2 “Heads” “Tails” 0.1 2.1% 2.1% 0.1% 0.1% 3 “Tails” “Heads” 13.6% 13.6% 0.0 4 “Tails” “Tails” -3σ -2σ -1σ 0 1σ 2σ 3σ

When we represent this probability in the When we increase the number of coin flips form of a bar chart, we get: to 4, the bar chart will then look like this: The normal distribution is a probability function that Probability of all outcomes for Probability of all outcomes for describe how values of a variable are distributed. It is 2 coin flips 4 coin flips symmetrical where most values lie in the middle region and the probabilities for values further from the average 0.5 0.4 taper off equally in both directions. Extreme values in 0.4 0.3 both tails of the distribution are extremely unlikely (e.g. 2 getting 1000 “Tails” in 1000 coin flips). 0.2 Σ l x - µ l 0.2 0.2 The normal distribution fits many natural phenomena = 0.1 (e.g. heights, IQ score, blood pressure), making it one of σ 0.1 the most important probability distributions in statistics. N 0 0 2 “Heads” 1 “Heads” & 1 “Tails” 2 “Tails” 4 “Heads” 3 “Heads” 2 “Heads” 1 “Heads” 4 “Tails” To see the normal distribution in action, you may want to & 1 “Tails” & 2 “Tails” & 3 “Tails” refer to this video: Probability Probability https://youtu.be/Kq7e6cj2nDw

8 STAR LECTURE 2020 STAR LECTURE 2020 9 Applications of Limitations of STATISTICS AND STATISTICS AND Probability Probability Probability is not knowing what will happen, but what is most likely to happen. Type l and Type ll error

Predicting volcanic eruptions Understanding the limitations of probability is You’re pregnant crucial to utilising it effectively. One such limitation is known as “Type I and type II error”. Type I error refers to giving a false positive (e.g. telling a patient he has COVID-19 while in reality, he does not have it) while type II error refers to By looking out for these signs, scientist can giving a false negative (e.g. telling a patient he increase their certainty of whether a does not have COVID-19 while in reality he volcano is near eruption and offer more does). accurate predictions. Type l error Such errors can have huge impact in areas like Mount Fuji, an (false positive) Ball touch event data the medical field. An example is given below: active stratovolcano in Japan You’re not Assuming there is an illness that affects 10 When scientists tell governments or the pregnant people out of every 100. The test for such an general public about the potential of a illness has an accuracy of 80% (20% of the time disaster hitting, they often use probabilities. the test result will be incorrect). For volcanos, there are various signs and symptoms scientists look out for prior to their eruption. Signs and symptoms include, but not limited to: Switzerland • Seismic activities: As magma exerts Type ll error pressure on the rocks beneath, the and Albania playing soccer (false negative) rocks crack and make a small Switzerland and Albania playing soccer, Albinfo, licensed under Creative Commons Attribution- ShareAlike 2.5 Generic earthquake Another application of statistics can be Test showing positive Test showing negative • Gas emissions: As magma nears the found in soccer. Soccer teams in the (with COVID-19) (without COVID-19) surface, the pressure it experiences Premier League employs mathematicians Healthy people = 90 18 72 decreases, allowing gases within the to improve the result of the team on the magma to escape playing field. One such method employed involves the use of ball touch event data. Sick people = 10 8 2 ª Ground deformation: Swelling of the During each game, approximately 2000 ball volcano indicate an accumulation of touch event happens, and all of these are After running this test with all 100 people, approximately 18 type I error will occur along magma near the surface recorded based on the position it occurs. with 2 type II error. From the table above, we can see that of all the tests showing positive, Analysing these data allow players to 18 out of 26 are actually healthy. This is the reason why the same medical test has to be run evaluate their performance after the game. multiple times on the same patient, to effectively reduce errors in the test. 10 STAR LECTURE 2020 STAR LECTURE 2020 11 Cognitive Prisoner’s Bias Dilemma

A cognitive bias is a systematic error in thinking that affects decisions and Game theory is the study of mathematical models of strategic interaction between rational judgements that people make. A demonstration was done in the 2019 Christmas players. An example of game theory is the prisoner’s dilemma. Lecture where audience were told to “pick a number, completely at random, between 1 and 10”. We would assume that approximately 10% of the audience In the prisoner’s dilemma, 2? criminals, A and B, are given 2 choices, either to confess or to would have picked each number. However, the result shows that barely anyone stay silent. They are not allowed contact with each other, thus A have no idea what choice picked the number 1 or 10 while the number 7 was most popular. did B make, and vice versa. The result of their choices are as follows:

• If A and B both choose to confess, each of them serves 2 years in prison • If A confesses, while B stays silent, A will be set free and B will serve 3 years in prison • If B confesses, while A stays silent, B will be set free and A will serve 3 years in prison • If A and B both choose to stay silent, both will each serve 1 year in prison

The above can be summarised in the following table:

B stays silent B confesses

A will get 1 year A will get 3 years A stays silent B will get 1 year B will be free A will be free A will get 2 years A confesses B will get 3 years B will get 2 years

From the table above, it is clear to us that the optimal choice would be for both A and B to stay silent, resulting in a total of 2 years served between both. The least optimal choice Number of people would be for both to confess, resulting in a total of 4 years served between both. who chose 7 was asked to stand up However, if we view the situation from the perspective of A, assuming B stays silent: Attempts by Mathematician Dr Hannah Fry at explaining this phenomenon in the 2019 Christmas Lecture includes the number 1 B stays silent B confesses and 10 being too extreme, 5 being “too much in the middle” and after more A stays silent A will get 1 year processes of elimination, ends up with the number 7. A confesses A will be free Perhaps you can try this out at your next gathering to see if it turns out to be true. From the table above, it seems clear that the choice which benefits A the most would be to confess if B chooses to stay silent. What if B chooses to confess?

12 STAR LECTURE 2020 STAR LECTURE 2020 13 Prisoner’s Dilemma Applications of Mathematics and B stays silent B confesses A stays silent A will get 3 years Algorithms A confesses A will get 2 years Computer Simulation Nowadays, movies are relying more and more on computer stimulation to produce those high intensity scenes where cities get destroyed. These computer stimulations that we The most beneficial choice would still be for enjoy so much are mathematical modelling performed on a computer, predicting the A to confess, resulting in A serving 2 years behaviour or outcome of a real-world scenario. rather than 3.

This perspective applies to B as well, regardless if A confesses or stays silent, the A real-life example of the prisoner’s most beneficial choice for B will be to dilemma can be found in the advertising confess. industry. Companies are able to save money if every company decide not to As a result, both A and B will choose to advertise their products. However, when confess, resulting in the least optimal this scenario happens, each company could situation. The prisoner’s dilemma explains advertise to gain an edge over their why two completely rational individuals competitors. This led to every company might not cooperate, even if it appears to investing money into advertisements. As a be in their best interests to do so. result, companies are choosing the irrational option of advertising over the Nuclear explosion test rational option of not advertising, even though the market share they end up in Nevada in 1953 obtaining may be the same. One of the first uses of computer stimulation occurs during the World War II. During the Manhattan Project, nuclear bombs were developed. Physical testing of these bombs could not be conducted widely. Thus, computer stimulation was brought in to model the process of nuclear ? detonation. 14 STAR LECTURE 2020 ? STAR LECTURE 2020 15 Applications of Machine Mathematics AND LEarning Algorithms Finding Organ Donors In 2012, Alvin Elliot Roth won the Nobel Memorial Prize in Economic Sciences. His ground-breaking work in matching kidney donors and recipients using computer algorithms allowed more patients to obtain a suitable kidney. Below is an explanation of how it works:

Imagine a donor, A , wanting to donate his 1 Potential Potential kidney to a recipient, A2. However, the kidney is not a match and the transplant will Living Donors Recipients not work. At the same time, another donor,

B1, wants to do the same for another

recipient, B2, but faces the same problem.

Fortunately, Donor A1 is a match for A self-learning Recipient B while Donor B is a match for 2 1 B A robot, iCub Recipient A . 1 2 2 A self-learning robot, iCub, Niccolo Caranti, In the past, computers relied on conventional licensed under Creative Commons Attribution-ShareAlike 4.0 International Trying to find such matches manually will programming. The approach of conventional programming be tedious and time-consuming. Luckily, involves feeding the computer a set of instructions for a computer algorithms can sieve through all defined set of scenarios. As a result, the computer is not the data and successfully match pair A -A 1 2 equipped to deal with anything outside of its to B -B allowing a 4-party (2 donors and 1 2, programmed scenarios. For simple tasks (e.g. playing 2 recipients) kidney exchange to be performed. Tic-Tac-Toe), it is possible to programme all possible scenarios. However, for more advanced tasks or even A B 1 2 complex games such as chess or go, it will be challenging for humans to manually input all necessary sequence of actions to make a decision. Furthermore, the sheer number of combinations will mean that the computer Visual representation of a will run out of memory before checking all the possibilities. 4-party kidney exchange In April 2008, a 12-party (6 donors and 6 This is where machine learning comes in. Instead of recipients) kidney exchange was performed. manually inputting all the sequences, machine learning allows the computer to learn how it can perform the tasks through trial-and-error, effectively writing its own algorithms to deal with any scenarios it faces.

16 STAR LECTURE 2020 STAR LECTURE 2020 17 p(BlA)p(A) p(B) Applications of p(AlB) = Machine learning Bayesian Machine Learning

For example, imagine you are teaching a computer to learn to flip a coin. One way (a non-Bayesian way) for the computer to learn is to flip the coin 10 times, and to record down the number of “Heads” and Generation and reconstruction of 3D shapes “Tails” it ends up with. However, with 10 from single or multi-view depth maps flips, it is unlikely for the result to show Generation and reconstruction of 3D shapes from single or multi-view depth maps, Arsi Warrior, licensed under Creative Commons exactly 5 “Heads” and 5 “Tails”. Perhaps the Attribution-ShareAlike 4.0 International result achieved is 7 “Heads” and 3 “Tails”. One of the most common applications of We know that the coin’s probability of machine learning is image recognition. landing on “Heads” to be . Using Bayesian Image recognition involves creating a machine learning, you can input this neural network that process individual pixel knowledge into the computer, allowing the of an image. Programmers “teach” these computer to come up with a more accurate networks by inputting as many pre-labelled model for flipping a coin. images, a dog for example, as they can, Thomas Bayes, allowing the programme to “learn” what formulator of Bayes’ theorem important features to look out for. The computer then develops a general idea of what an image a dog should have in it. Bayesian machine learning enables the When a new image is introduced to the programmer to input prior knowledge into computer, it compares this image to every the machine, regardless of what the data picture of a dog it has ever seen. If the new might tell us. This is particularly useful when image has a high level of feature similarity dealing with small data sets that might result in error in the algorithms that the to the computer’s internal representation of machine writes by itself. “dogs”, the computer will then declare that it is a dog.

18 STAR LECTURE 2020 STAR LECTURE 2020 1919 Limitation of Potential Facial Recognition Pitfalls

Other than telling if a picture is a face, Facebook’s facial recognition software go Privacy a step further and is capable of accurately identifying whom the face belongs to. If you use Facebook, you might realize that sometime, you get notified when With our increase dependency in the internet, our online privacy becomes an important someone uploads a photo of you even when you are not tagged in it. Despite issue we must address. Every time we use the internet, we are exposing a small part about these immerse capabilities, the software is not infallible. ourselves to others. One way this can happen exists in the form of a cookie.

1. The browser requests a web page

2. The server sends the page and the cookie

W The cookie W eb br o Hello World! eb br o w w ser ser

3. The browser requests another page from the same server

The cookie A possible interaction between a web browser and a web server involving a HTTP cookie Breaking key features of the face like the edge of the eyes and mouth allows a person to HTTP cookies, often referred simply to as cookies, are small files stored on a user’s computer containing data evade facial recognition software specific to either the user or the website visited. Cookies Facial recognition software works by picking out key can be accessed by the web server which then allows the features of the face like the eyes, the nose and the server to raise suggestions tailored to the user’s mouth. To pick out these features, it depends on the preference. When used correctly, cookie allows a more contrast present (e.g. between the eye and the eye relevant web browsing experience. lashes). This means that if we are able to break these contrasts, we are able to fool the software into ignoring However, one drawback of cookies is that they enable the face it is currently looking at. data mining. During data mining, users’ data are unknowingly examined in order to generate new To get started with learning machine learning for information. This action of data mining raises questions yourself, check out: regarding privacy. Information about the user which the https://machinelearningforkids.co.uk user did not intend to make public may be exposed as a result. 2200 STAR LECTURE 2020 STAR LECTURE 2020 21 No. Ki12t Potential Pitfalls Science Debate Kit: Sorting fact from fiction Misinformation is a growing concern in this internet age, where anyone can put Self-driving cars pieces of false or inaccurate information online. The impact of misinformation can vary from a prank to even malicious content such as hoaxes. “Keep these kits coming please!” An extreme tool that has been employed to propagate misinformation is the programme known as deepfake. Photo manipulation (e.g. adding or removing details in still images) has been around for an extended period of time. However, For in-depth online resources on this debate go to: cars.imascientist.org.uk deepfake makes use of machine learning to manipulate or generate visual or audio content to make highly realistic products. Essentially, with the use of deepfake, you can make it seems like someone is acting or speaking in a manner Debate Kit: Self-driving cars in which he/she never did. Should our town centre be for self-driving cars only? A structured practice debate on a controversial topic To understand more about deepfake, you may want to refer to this video: The different ‘rounds’ of the debate help students think through the issues and reconsider their opinions. https://youtu.be/gLoI9hAX9dw The structure also shows them how to build a discussion and back up their opinions with facts.

You can use all eight characters, Characters or fewer, as you wish. Yes No

The minimum is the four essential • Dara Attar – Tech entrepreneur • Michael Owuo - Epidemiologist and cyclist characters (in bold), this gives • Fiona Campbell - City planner • Lisa Dixon - Taxi driver two for and two against. • Bethany Fisher - Sight impaired person • Rachel Fong - Rural mum • Kazik Majewska - Commuter • Dave Lyons - Warehouse operative

Facilitation tips • Ensure pupils know there is no right or wrong answer. • Be observant of ones who want to speak and are not getting a chance. Designed for KS4 but can be used with ages 11-18. • Encourage students to give a reason for their opinions. For groups who may need extra support you can put the following prompt sentences upon the board: “I think we should/shouldn’t make the town centre for self-driving cars only because...” “I think ……………… is the most important point to think about.” Learning notes Learning Other learning outcomes: Curriculum points covered: SUMMARY objectives: • Consider social, ethical and Thinking scientifically • To practise discussing factual issues in an integrated • Evaluating the implications of technological and debating issues and way applications of science While a large part of our lives is governed by expressing an opinion • Think about different points of • Developing an argument mathematics (probabilities, algorithms and view • Reflecting on modern developments in science models), it is of utmost importance not to • Learn to back up their opinions develop an over-reliance on it. We should with facts Substantive strive to understand how it works and at the • Consider practical aspects of motions and forces, particularly driver and vehicle interactions and road safety. same time, be aware of its limitations.

“Particularly like the format plus 22 STAR LECTURE 2020 the accuracy of the scientific information” This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/4.0/ Rachel Fong - There are big questioDarans a Attarbout e –x actly how the driverless Michaelcar Owuo - Many of these technLisaolog ieDixons are -se nsing objects around the Rural mum Tech entrepreneur Epidemiologist Taxi driver future is going to work. At the moment development is maandinly cyclist car in different, complementary, ways: for example, cameras I’ve got two children under 3 and Ib eing driven by the luSelf-drivingxury c carsar arem thear future.ket And- th toe vehicles are expensive. don’t work well in lowI’ve lbeenigh drivingt, rad a taxia forr w teno years,rks best for metal objects, be honest, they are just cool! We should be leading in and I really love my job. I’m ‘Lisa’s Lady live in a rural village. I often come Aintos early adopters with large disposable incomes start usiI’mng a scientist who studies what makesultr peopleaso illn andic sensors only work over small distances but can ‘see town to go to the shops, go to soft play, visit friends this technology. There are different levels of how to make the health of the whole population better. Cabs’. Lots of women feel happier with a female taxi or family. It’s a nightmare getting two d kidsriv intoer theirles cars cars moreautonomous - wha cart ha technologyppe n- froms t levelo o 1,t whichher isr oad-users?‘Active travel’ - walking and cyclingt h- makesrou gsoh much’ som e objecdriver,ts. especially travelling on their own, late at night. seats and all their stuff, plus the shopping, in the boot. things like cruise control, to 5, which is a car that sense, especially for short journeys. It makes us fitter, I’m company for people. One old lady books me every I would just stop coming into town altogether if I couldn’t doesn’t need a driver at all and can drive itself in any healthier and happier, and it doesn’t cause pollution and week and takes flowers to her husband’s grave. She tells drive in. I’m happy with things the way they are, we don’t situation. Level 4 is a car congestion. It’s also good for communities, we say hello to me about her life and when she was young. A computer always need ‘progress’. Some people suggethats tdrives th aitselft w- a etrue w driverlessill ne carx -t but h thatav onlye operatesrobo t taxis - tpeople,he we pop into local shops. Fewer car journeys overall algorithm wouldn’t be the same for her. in a particular, controlled area - like our town centre. We need are better for us, better for our communities, and better for driver is 70% of the cost of a taxi, so robot taxis could be the planet! Fact: Self-driving cars are far more expensive than a lot of practice at level 4 in order to get to level 5 and truly Fact: One study suggests that 1.2 million driving jobs could ordinary cars, and likely to remain soc forh ae longap time.er. If people wunlockho thisc atechnology.n afford it switch to robot taxis, tFact:ha Int the 1970s the Netherlands invested in cycling be lost in the UK - for example, taxi drivers, bus drivers, infrastructure and promoting cycling. Now over ¼ of all delivery drivers. Issue: This seems like it’s going to make a lot of people’s Fact: A government report says that self-driving cars could trips are made by bicycle - compared to 2% in the UK. could mean only the poorest will still use buses. This could Issue: If we get rid of human beings, we lose the human lives LESS convenient, just for the sake of shiny new tech. be worth an extra £52 billion to the UK economy by 2035. Issue: We should prioritise making our towns more touch. Life shouldn’t only be about what’s most efficient. Who is going to be able tole affordad to tdriveo l eintos s funding for public transport, and ‘transport deserts’, Question: Issue: Self-driving cars are coming, whether we like it or walkable and bikeable - not still centre them around the car! the city centre? Do we want only rich people to have Question: No computer system is unhackable. in poorer areas. not. Can you hand on heart say that self-driving cars convenience? Question: My customers feel safe with me, will they be as safe will REDUCE the number of journeys people make by car? Question: Shouldn’t UK business get in there, get involved, with a computer programme Cyber security arouandnd reap d theriv benefits?erless cars raises a lot of issues - we controlling where they go? could have written a whole kit about that, but then we’d have missed out many other issues. But your students may want to think about what could happen if your driverless car computer was hacked? And whether the security services should be able All the facts in this kit have been researched. References can be found online at: to override the computer on your driverless car? (cars.imascientist.org.uk) Dave Lyons - Fiona Campbell - Kazik Majewska - Bethany Fisher - Warehouse City planner With special thanks to MichaelSight Talbo impairedt, Head of Strategy at Zenzic, a organisation Important technologies that make self-driving cars posCommutersible funded by the UK government and industry to help guide the development of Operative • Cameras to sense what is around the car. person The UK population is increasing, and self-driving technology. Prof Nick Reed of Reed Mobility, Professor Natasha Merat, I commute for two hours each • GPS to determine peoplethe makeca rmore’s lcaro cjourneysatio nall. the time. I think self-driving cars are a great idea, but I think this is day. I have a specialist job and I can’tInstitute get a job anywherefor Transport Studies,I am sight University impaired. It's of not Leeds, severe and Professor I can see Andrew Maynard, Our towns and cities need to accommodate that. the wrong way to go about it. Human error causes about closer. We live in the village my wife grew up in, and her things, but it is very blurry - like a fully sighted person • Altimeters, gyroscSelf-drivingopes, carsan couldd t ameanch ay systemme ofte pooledrs t carso keep track of the Risk Innovation Lab, Arizona State University, Dr Jack Stilgoe of UCL, principle 75% of car accidents, which is why I’m a fan of parents help with childcare for our young twins. We can’t looking at the world through a piece of fabric. Of (like robot taxis) and so far fewer cars would be needed to investigator on the Driverless Futures project and author of ‘Who’s Driving self-driving cars. Computers don’t drinkp andre drive,cis gete position of the car. move to be nearer work. There’s no bus route I could use, but course I can’t safely drive a car. Sometimes (especially drive everyone around. These cars could be connected distracted or drive when they are tired. But we should start I hate spending so much time driving eachInnovation? day. New Technologiesat night) it’sand even the very Collaborativehard for me to get a bus State or walk’ and Perry Walker of and talking to each other, and reduce congestion. It’s easier with lorries carrying freight. They do a •lot Rof miles,ada oftenr, aon system that uses radio waves to work out the range, somewhere. And taxis are expensive. Driverless cars for level 4 self-driving cars to operate in the centre if they In 2016 3.7 million people in theT UKalk commuted Shop, for dialogue project. rural roads, and they follow set routes - perfect for level 4 Fact: would transform my life. angle and velocitydon’t of haveob toje dealc twiths. normal cars too. two hours or more each working day. That’s 32% more technology. than in 2010. Fact: There are 350,000 people registered blind or • Lidar, a sensing syFact:ste Today’sm w carshi arech only fo beingllo usedws 3.5% th eof the p time.rin ciples of radar, The kit has been produced by the I’m a Scientist team and funded by The Royal Fact: Most people who are killed in car accidents aren’t in partially sighted in the UK, and 2 million living with sight They are parked 96.5% of the time. Issue: A self-driving car can use cameras or lasers to city centres (where traffic moves slowly), they are killed on Institution, Lloyds Registerloss Foundation that affects their livesand (e.g. Institute they are not of able Physics. to drive). but uses light from a laser. Detects what is around the cdetectar. its surroundings, lots of computing power, and rural roads. • Ultrasonic sensorsIssue: de Connectedtect o autonomousbjects cars n ecana maker th thousandse ve hicle, andadvanced the AI to drive the car more safely than I can, while I Issue: Driverless cars mean that many people would be of calculations every minute and be safer and more efficient do something more interesting! Issue: Computers are good at ‘bottom up’ processing (e.g. able to make journeys independently, for the first time. than humans. seeing edges of objects), but they aren’tc goodar ’ats ‘top m ovements. This includes people with impaired sight and some other Question: I don’t wash clothes by hand, I use a washing down’ processing - e.g. knowing what a cat is and what it This work is licensed underdisabilities, the Creative teenagers Commons and old people. Attribution-Non • Advanced AI/comQuestion:puter Doessy sit maketem senses to for integrate all this datamachine., Why would we still want to do boring jobs that might do. City centres don’t play to their strengths. Commercial-ShareAlike 4.0 International License. To view a copy of everyone to have their own car, machines can do for us? Question: Not everyone can drive, maps and knowledbutg onlye auseb ito sou littlet r ofo thea dtime? signs, driver behaviour, etc. Question: Should we give up our this license, visit http://creativecommons.org/licenses/by-nc-sa/4/04.walk or cycle. Don’t I deserve the city centres without getting the real same freedoms that other people benefits of driverless car technology? take for granted? K it N o 12 Main Activity: 35MaMa minininu ActActtesiivv.iity:ty: 3535 mminuinutestes.. Teacher Notes 1) Spli t student1)s1) in SpliSpli to astt st stmanududeeynn groupttss ininttoos asasas manmancharayy groupctgroupers youss aas swan ccharaharat toctct ererss youyou wanwantt ttoo cover. covecoverr.. Question: Should our town centre 2) Give them their2)2) cGi Giharaveve tctthemhemer c ardttheirheirs –cchara haraone ctctpererer cgroupcardardss , – –and oneone gi perperve groupgroup,, andand gigivvee be for self-driving cars only? them a few minutthemeshem to a aread ffewew tminuminuhemt tesovees ttoro. readread tthemhem oveoverr.. 3) Get one studen3)3)t G Gine eeatt oneonech group ststudenuden tott ininrea eaeadc couhh groupgroupt their t too rearea sectdd ououiott tntheirheir to the rest of thet toclasso tthehe . rerestst ooff tthehe classclass..

Lesson plan What are the claWsWsha’has tinit arearetial t thethehough claclasstsss’?s’s iniIsini tttialhereial tthoughhough one potsts??s IsitIsion ttherehere the y oneone popossititionion ttheheyy identify with or rejecidenidentt?tififyy wiwitthh oror rejecrejectt?? The different ‘rounds’ of the debate help students think 4) Take it in turn 4)t4)o Trea Takeaked ioutit inint ttturnheirurn t tofaco reareat. dDoed ououstt ittheir heirchange facfactt. . t DoeheDoe wass iytit cchangehange tthehe wawayy through the issues and reconsider their opinions. The structure they think? ttheyhey tthinhinkk?? also shows them how to build a discussion and back up their 5) Read the issu5)5)e . Rea ReaAnyd ddi tthefheferen iisssstu ufeeelinge.. A Anynys di?diffferenerentt ffeelingeelingss?? opinions with facts. 6) Each team ask66s)) EtEhaeacicrh hq ttueeaeamsmt i aoasnskk tsso t ththheeeiri r ch qquuaeersastctiiotoennr totoof ttthheeei chchr caahrroiaaccttee.rr ooff tthheieirr cchhoioiccee.. Starter: 5 minutes. Support: To helpSupport:Support: students TT oyouo helphelp can studentsstudents put the youfollowingyou cancan put putprompt thethe followingfollowing promptprompt How do they think the invention of cars affected the world? sentences up onsentencessentences the board: upup onon thethe board:board: Prompt, if necessary, them to think about: the environment, “ I think our town “cen“ I I tthinkhinktre oursourhould ttownown/shouldn’ cencenttrere t s sbehouldhould for /ss/selhouldn’houldn’f-drivingtt bebe c ar fforors sselelff-dri-drivvinging ccararss how cities and towns are laid out, how people live, where they only because...”onlyonly becausebecause...... ”” live, where they work, communities, families. “I think …………“……“II tthinkhink is …t…he………… mo…st… …………impor isist an tthehet poin momosttst t oimporimpor thinkttan anaboutt poinpoint.”tt ttoo tthinhinkk abouabout.t.”” Do they think people foresaw all those effects when cars Plenary: 10 minuPlPlteeennsary:ary: 1010 mminuinutetess Vote for which poVVsooitteione fforor they whichwhich agree popos swiititionionth mo ttheyheyst agree(iagreef there wiwi tisthh one) momostst. (i(iff ttherehere isis one)one).. Engineers are developing driverless car technology right Why? Which argumenWWhy?hy? t WsW werehihicchh argumen argumenthe mostt tspersuasis werewere ttheheve mo?mostst persuasipersuasivvee?? now. Today we’re going to think a bit about what that might Note – Pupils canNN osottayee – –in PP upilrolesupilss callcanan t he ssttay ayway inin rolestroleshrough allall tdebathehe waywayte, or tthroughhrough only debadebattee,, oror onlonlyy mean in the future. a chance to expresaa cschance hancetheir own ttoo expresexpres opinionss t theiraheirt the ownown end opinionopinion and in atahett tthe heplenar endend y andand. inin tthehe plenarplenaryy..

help to have themhelphelp st ar ttoot havebhavey di sc tthemhemuss ing ststarar tthet bby yque didiscscstuionussss ingingand tthe/heor quetqueheistrst ionion andand//oror ttheiheirr character’s positcioncharacharac in pairtteerr’s’s,s posi posiandtt ionionthen inin c pairpairompares,s, andand no tthenthenes in ccompare omparefours. nonotteess inin ffourourss.. Designed for KS4. These debate kits have been used with ages 11-18. They’ve then hadTT hehechanyy’ve’vec e tthen hento rehear hadhad ccshanhane scomecee ttoo o rehearrehearf whats steehe ssomeyome wan ootf f whawhatt ttheheyy wanwantt to say before hatvtooing ssayay to be bedoffore oreit in ha hafronvvingingt o tfto ot he dodo whole iitt inin ffronron clatt ss ooff . tthehe wholewhole cclalassss.. K it N o Many of these technologies are sensing objects around the There are big questions about exactly how the driverless car car in different, complementary, ways: for example, cameras Teacfuturhe eris g Noinogte tos work. At the moment development is mainly12 don’t work well in low light, radar works best for metal objects, being driven by the luxury car market - the vehicles are expensive. ultrasonic sensors only work over small distances but can ‘see Question: As early ad oShouldpters wi thour larg towne dispos centreable inc o mes start using through’ some objects. driverless cars more - what happens to other road-users? be for self-driving cars only? Some people suggest that we will next have robot taxis - the driver is 70% of the cost of a taxi, so robot taxis could be cheaper. If people who can afford it switch to robot taxis, that Lesscouldon me apln oannly the poorest will still use buses. This could Thele adidff tereno lests ‘rounds funding’ o ffo trhe pu debablic trtea nhelpspo rst,t udenand ‘tstr anthinspok rt deserts’, throughin poo trheer arissueseas. and reconsider their opinions. The structure also shows them how to build a discussion and back up their opinionCybesr swietchu fraitcty sa.round driverless cars raises a lot of issues - we could have written a whole kit about that, but then we’d have Starter:misse d5 omuinut matesny. other issues. But your students may want to Howthi ndok a tbheyou tt hinkwha tt heco uinldv enhatpionpe no fi fc yarsou ra dffrecivteedrle tshes c worldar co?m puter All the facts in this kit have been researched. References can be found online at: Prompwas t,h aifc nekecdes? Asarndy ,w themheth etor tthhinke se aboucurityt: setherv environmenices shouldt ,b e able (cars.imascientist.org.uk) howto ocivtieserr idande th teowns com areput elaidr o nou yot,u howr dri vpeopleerless livecar,? where they live, where they work, communities, families. With special thanks to Michael Talbot, Head of Strategy at Zenzic, a organisation funded by the UK government and industry to help guide the development of Important technologies that make self-driving cars possible self-driving technology. Prof Nick Reed of Reed Mobility, Professor Natasha Merat, Do• tCheyam tehinrask tpeopleo sens efore whsaawt is all ar othoseund tehfefe cctasr. when cars Institute for Transport Studies, University of Leeds, Professor Andrew Maynard, • GPS to determine the car’s location. Risk Innovation Lab, Arizona State University, Dr Jack Stilgoe of UCL, principle • Altimeters, gyroscopes, and tachymeters to keep track of the investigator on the Driverless Futures project and author of ‘Who’s Driving Engineerprecisse are po sdevelopingition of the driverlescar. s car technology right Innovation? New Technologies and the Collaborative State’ and Perry Walker of • Radar, a system that uses radio waves to work out the range, Talk Shop, dialogue project. now. Today we’re going to think a bit about what that might meanang inle t hean dfu vtureelo.city of objects. The kit has been produced by the I’m a Scientist team and funded by The Royal • Lidar, a sensing system which follows the principles of radar, Institution, Lloyds Register Foundation and Institute of Physics. but uses light from a laser. Detects what is around the car. • Ultrasonic sensors detect objects near the vehicle, and the car’s movements. This work is licensed under the Creative Commons Attribution-Non • Advanced AI/computer systems to integrate all this data, Commercial-ShareAlike 4.0 International License. To view a copy of Desmignapesd afonr dK Sk4n. oThweseled dgeeb ateab okiutst road signs, driver behaviour, etc. this license, visit http://creativecommons.org/licenses/by-nc-sa/4/04. have been used with ages 11-18. BackgroundBackBack nogtegrrsoundound nonotetess Key issues KeyKey iissssuueess There are big questions about exactly how the driverless car future is going to work. At the moment development is mainly Self-driving carSsSeellf-f-ddrriivvinging carcarss we get to level 3w woere 4 gg.e eHtt tutoom lleanevveesl l a33r oeobr rn e 44oi.n. t H gHg uodumromivdanane ansts b s aaywr reteiht c nenho oliutnt x ggguo o royod dc aartt msswwairittkccehhti in-n gtgh e vehicles are expensive. Self-driving carsS S(eaelfllfs--odd rcriivavilinlneggd c caaaurrstso (n(aaomllssooo ccuaasll leveded h aaicuultetoosnn)om omareoo uvuses h vviecehlhesiiccl leess)) aarree vveehhiaiccltlestesen tion and theaaytt ttaeernnetiotio nnon t a agnnoddo tdthh eaeyty A p aasarr eye i nangrool tyta ggatotodeoonddptio taaenttr pspfo aawryy iiatinhn glo gla anargtgttte en ndtiotioisnposn ffoorar abla loelo ninnggc omes start using which can senseww thhieicc hhe nccvaainrno ssneemnnsesene t t thahereo euennnvdvi irrtoheonnmme eannntt d aa rtroroauuvnneddl tstheheafemmly a anndd ttrraavveel l ssaftafimeelelyy w hen they mottiimmseet l wyw heaherennn t’thth eneyey emomod desrsditvt.lly eyI fr a alyerereosunsn ’ ’tact nreanerensee’ dtmod eengeddr.e. Ia If-f g yyweoodhuu,a ata reharennp’’tpt engeengnsa atgoge eoddt,h, er road-users? without human inwwpiituthhto o(uoutrt huwhuitmmh aalenns isinn phupuutmt ((oaornr w winiitpthhu ltlee tsshssa hunhu amm ‘naaonnr iminnpapulu’t t c ttahhraa)n.n aa ‘n‘noorrmmayalol’ ’ ucc aagrre)).t. distracteyydoo uau f gtgeeert t a ddbiisostturrata c2ct0tee dmid aanfftuteetrer asab.b Ioof uuthtt e22 00c amimir nsnuduutteedsse.. n IIflf y tt hhee ccaarr ssududddeennllyy a lerts you, ‘help,a awleleerrt tsars yeyo oauub,, o ‘‘hhueet llptpo,, wcwSoreae s armarhe!e’ aipeoat bbtaookupuetlte s tt o ost iu cmcgrraegas esfhohs!r!t’ ’y tiithot tautaat k tkwoee ess twtiimmille en fefooxrrt yyhooauuv ettoo r obot taxis - the referred to as sixrre,e ftfoee rrirnreecddlu ttdooe aa sst sasinixxd,, a ttorod iin ncccalluru dtdeeec shstntaaolnndodagaryrdd) . c caarr tteecchhnnololooggyy)).. return your attenrrtioeettunur rntno y ytohoueur r r aoatattteden natiotionddnrni vw ttoeoo r tr thikhse e o7 rur0oot%a awdd ho aaf ntn hdidse w w gcooorirsknkt g ooof uou tnat w.w thahaxatit, iiss o gg oroobiinngog t o otnan.x. is could be cheaper. If people who can afford it switch to robot taxis, that StandaLLrevdev ceeal lr 0 0technoSSltotaagnnyd,d atharrded c hcauarmr t teaecnch hdnnoooelolsog gayyl,l , th ththee h dhuruimvminaagnn d dooeess a all lth thee d drriviviningg Working out infoWrWmooarrktiokiinnngg a oobuuott u iintn ftfohorermm ceaoantioutiovldirnn o m anabemboaoeununt t ot t thnahelreyo eutehnnnevdv i irpryooonnoummr eeesnnt ttw aairlrol osuutninldld u yysooeuu b uses. This could A singlLeLev eveleeelm l 11ent of AthA s esi nidngrgilvelei ne eglel epmmroeecnnet tso osf f th hthaees d dbrreiviveininng gt ap pkrroeocnce eossvsse h rh a ass b beeeenn t taakkeenn o ovvederer pends on, brodadedeplpyenen, twoddss okoninn,, d bbsrr ooofaa dledprllyayo,,d c tt woewotos s l keekisinsns:d-d fssu ofnofd priprngoo ccfoeerss spseuesbs:l:-i- c transport, and ‘transport deserts’, by the ‘ch‘haaarn n-d dess. g oo.n nc’’ruibsbyey th cthoeen c tcraoarlr ,- - oe er. g.lga. .nc cerru-ukisiesee pc co oannststrroiosl,lt ,.o orr l alannee--kkeeeepp a assssisistt.. 1. Bottom-up pro11c.. e B Bsoostetttosom- m-(fouurp pe prxpraoomccipenels espss,oe einoss r (ve(fiforso ruaarr eeexlx aprasmmo. pcplelees,, si iinnn vgvi,is suaual l prproocceesssisinngg,, CompuLLteveevrsee lc l 2a2n takeCC ooovmpmper umutteuerlrtiss pc claean nf ut tanakcketei ono ovvsee rfr m ommuul tilttihppelel ed f rfuiuvnneccrt,ti onionss f frroomm t thhee d drriviveerr,, ‘hands off’ bringin‘gh‘h ataonngddessth ooefffrf’ ’i nfbobrrinminggainitnigog nt to ofgrgoeemthth evearr ri ninoffouorsrm msaeatntioisonon rf sfrro otmom cv voaanrritoiroouuls sth s see nnssoorrss t too c coonnttrrool ldth theeet e cting edges)ddeetteeccttiinngg eeddggeess)) speed -a- bnbudut t d nnioroett c tions soppfe etheeded a canandrd md diurirecechct tioiofn nt ho oef f th tthimee ec c.a aHrr m ummuucanchh so o fs ft thilhle ene t timiemdee . .H Huummananss s s2ttil.ill l Tne neoeped-d d own pro2c2.e. Ts Tosopep-s-d d(ofoowwrn ne prxpraoomccpeCelsesyss,be eyessor (us(ff oeokrcnr euoexrwxiata ymwm ahppraloleetu, , an yy doco uaud t rk kinvsneo orwwle wswshh acatat aras c craatist iises s a lot of issues - we to be ‘hlilatitererndaaslly lyo!!n’ at attolol tb ibmee e‘ h‘sha a-n nmddsos n oointno’ ’a raitn ta gal lth lt tiemim deersisv - i- nm mgo oannnitidtoo rrrieninaggd th yth eteo d drriviviningg a anndd r reeaaddyay t ntoo d how it mighta abnneddh hahovowew) iitt mmiigghhctt o bbueelhdha ahvvaeev))e written a whole kit about that, but then we’d have interve n e in an emeirningtteernrvcveyen.nee i nin a ann e emmeerrggeennccyy.. missed out many other issues. But your students may want to Not exLaLevcevtlyee la l 2 2‘+l+evel’,N Nbouott t e tehxxiasac citlstly yw a ah ‘ erl‘eleevve etlo’l,’d ,b abuyut’ tst th hmisiso i sist wawhdherveraene ct toeodda ayy’s’s m moosst ta addvvaanncceCeddu rrently, compuCCtueurrrsre eannrttellyy ,v, cecoroymm gppouuottetdehr risans t ka a1 rare,e b b vovueeutr rtny yw o ggtho ovaooetd drc y aoa tgut 1ol1d,o, bhadbu uatpt t np n2oeo.tn t v viefe ryryoy uggroo ododrdi v aaettr 2l2e..s s car computer vehicles are up to. Tvvheeehhyici calelreess a barereye ou unppd t toloev. .T Tehhl e2ey y -a athrree b yb eaeyryoeon ninddt el evlevgreaelt li2 n2 g- - ththeeyy a arree i nintteeggrraattiningg was hacked? And whether the security services should be able more sensor data, inmmcolourrede is nsegen nmssooornr d idtaoatrtaian, ,gi ni ntchcleulu ddirninivgge m rm. oBonnuitito otrhrineingyg at thrhee nd droritviv eerr. .B Buut tt thheeyTy a harree yn n odott o n’t have TT‘chhoeemyy mo ddoonn’ ’tst hehanavsvee ’ ‘ c‘acoonmmd momocannn’ t s smeenanskseee’ ’i ananfneddre ccnaacnne’’tt s mm, faaorkke e iinnffeererenncceess,, fforor yet at level 3. yyeet ta at tl elevveel l3 3.. example, to predeeixcxata mwmhppalleet ,,a tt oope ppdrreedsditicrcttiota w wnoh vhmaeatrit gr aaihd petepe d tdohde.e Ws scttreoriia mawnnp i lmulm ntieigegrheh otdt dn dao oy ..o WWure ed wrwiiville l nrnleeesedsd aca a r? The drLiLveveevre ecl al 33n safelTyTh hteue rd ndr rithviveeierr rc caaatntne s nsataifofeenlyl yt ot tu uorrntnh th ethrei eitrhir a natgttesen nftotioiro nmn t utooc o hot thheerr t thihinnggss f foorr lm omtu uofcchh a dvances ilnloo tAt ofofI t aeadcdhvvannanolccoeegssy i intno A AgIIe ttee dcchrhinvneololroloeggsyys ttcooa ggrese tt odd rlrieivvveeerrllle e5ss.ss ccaarrss ttoo lleevveel l 55.. of the t‘ie‘meyyee.s sT oohffeff’ ’vehoiocf flt ethh ecea t tinmim heea. .nT Tdhhlee bv vereahhkiciicnlelge c icnaa nan nh h aeanmnddelelreg b ebrnraackkyini,n gbg ui nitn a ann e emmeerrggeennccyy, ,b buutt Important technologies that make self-driving cars possible the driver must be avththaeeil d abdrrivlievee trro m minuutsestr tb vbeen av ave awailitilababhinlel ea t tolio mi ninittteeerdrvv eteinmneee w ,w iitfit hthhinien aa li limmititeedd t timimee, ,i fi ft thhee • Cameras to sense what is around the car. vehicle alerts them.vveehhiciclele a alelerrttss t thheemm.. Human drivers AHHRuuEmm gaaonno ddrr iaivvete trrossp AA-dRRoEEw gngo oproodod c aaett s ttosioppn--gdd,o oawwnnnd prprcaooncc eesssisinngg,, aanndd ccaann cope with many cscopitopueae tio wwinitthsh mtmhaaantn ydy r ssivit•iteu uGraaletioPtiosSsnn tscso a t thdrhasea twt e ddrormirinivv’etne rberlle ets hsasesbl cceaa rtrso’ss w wlooocnna’’ttiot bbene . aablblee ttoo The veLhLevievcleeel li 4s4 totallyTT hreheesp v veonehhiscicilbelel ei si sf ot toort tadalrliylvy ire nregspsp aonnonds sthibibelel eh f ufoomrr d adrnrivi vdininrgigv ea arnn dd th thee h huummaann d drriviveerr can lea‘m‘vemini ntdhdses odofrffifv’’erc’csa asnne l aeletaa, veover t sthhlee ed dprr.i viBveeurr’ts’,s ths seeaa vtt,e ,o horir cs slleele ecepap.n .B Bounuttl,y ,th th ee v veehhiciclele c caann o ocnnoplyly e with for a loccnopopg etei m wwieitth h( i ffo oerrv aea r lolo).n nL•g ge A tvtlitimemiml e3ee (a(tieiffn reedsvv ,e4 egr rc)y).ar. oLrLsee cvmvopeeela l 3y3 s ana,ne nadeddn 4d4 ctcaaacrrhssy mmaaeyyt e nnreseed edto keep track of the operate in geofenceodop paererreatataese ,i niwn g hgeeeoroeffe evnneccreyed dd a eartreaeaialses,d ,w wmhhaeeprrepe iv nveger ryiys d d eettaailieledd m maappppiningg ui sis to change theuu srso tatood c cehahanvnnigrgoeen tmthheee n rroto pa-a rdmde ecaenisknvievniir grpooo nensmdmigtioeeennsntt -cof- le mm taaharkeeki inrcn,g ag r e.eddggeess ccleleaarreerr,, available, lane markaviavngaasilil abaabnldele ,r ,lo alanndee em mdagarerkksini naggrses a canlnedda r ro o(aead.gd e. ecddigtgyee scs ea anrreter ec cslel,e aarr ( (ee.g.g. .c citityy banc ceennttrnreeisns,, g pedestrbanibanansnn, iinbngicg y pepecldedese,ss tbtrruiianang• gs Rsie,,ad bsb,icica iycry,ce cal lece srsye,, sa bbtmueugm gvgg anietiehsas,…t, iuiccseee ccsrr ereaadmmio v vwananasvs…e…s to work out the range, or an airport car parokor)r ,a aonnr a uainrirpdpoeorrrt tsc cpaaerr cp piaalr rkck)i),r ,co ourr mu unsndtdaeenrrc s esppse e(cecia.iagl .lc cairi rc cuummssttaanncceess ( (ee.g.g. . aa angle and velocity of objects. alert the driver, and asalaelefrert tlth yth epe ad drrkriv ivtheerer, ,ma ansnded ls vsaeafsfe eilfyl yt hp peaayrrk kd th othnee’mtm rsesespelvlveoesns di fi. ft thheeyy d doonn’t’ tr reespMspoonsndtd ..driverless cMMaoross tat ddrreri ivvbeeerrilnleegsss sd cecava•re rs Lslope i adarareerd , b b aaee niseinndgg n t desdiesnetvgveee dslopelope yins tpeddlma aca nwesnddh titceehss ttfeeoddll o iinwn pspl latahcceeses p r inciples of radar, The veLLhevevicelel l 5d5oesn’TtT hnehee e vvdee hahi cihcluelem ddaoonee sdsnnr’it’vt neener, eeevdd e aar , h hauunmmda acnna dndr ri viveerr,, eveveerr,, aanndd ccaann l i ke Palo Alto, Calliiklkifeeo PrPnaaialloo, wAAlhlttoeo,r, e CC raaolliadiffoobrsrunn tia iaaur,e, s w wewihshe delirergee h a rtrno ofdadrado smstsr a aarirg eelah wit,wise ddre.e D aanendtde scstttrsraa wiigghhatt,t, is around the car. correc‘tls‘sytte eaeenrridnin gsg a felycc ocororrrenectcrtlotlyly i atasnneddl f ssianaf feevelylye crcyoo ncntitrrocoul l imtitsseetlaflf n inicn e evev theearryty a cc irirccuummssttaannccee a tthhnaadtt aala id out on aaa gnnrddi d lala piidad t otoeuurtnt o,o nan n aad g g•tr hri Uidedl rt ppreaa itsstttoe elnrirtnintcl,e, as amnendnidx s tethohdere srre oed a ieisdst eluliitcttstltle eo mbmjieixxceetdsd rnrooeaaaddr utuhssee vehicle, and the wwheheeel l car’s movements. fete, aonopdpt tifoionlnloaawil’l’ ng fthfeteteee ,h, aanndd fsfooigllolnowiawilsnn gog f th thae ev hohalaunnddt e sseigirg.n nWaalisllsl oroeff q aau vvirooelul unntteeeerr.. WWilil l rr(eeiq.qeuu.i ricreey clists and ((pei.i.ee.d. cecysyctcrlilanissttsss a)a.n nAddn pedpe wddehesestrtreriian anthses)) .w. AAennaddth wwehrh eiesrr eue s tthuahee ll wwye eaatthheerr iiss uussuauallllyy very advanced AI. vDveeifrfryey r aaeddnvvta aennxccpeeddr t AsA IpI.. rDDeidfiffifecertre etnhntits e ewxxpilpele brrtetss av pprreaedildabicictt l tethh isis wwilill l bbee avavdarayililab abanleled sunny. Hddarryvy i naagnn ddth sseuu nsnannymy.. e HH taae•vv ci Ainhndgngv ol tanthhoeegc syes adawm moAereIk / tc teieonccm hthhnpneoluol ctoeoegrgn yyst r ywweso otemrrkk iinsn ttothhe ei n ccteengntrtrateree all this data, within 10-30 years, wswoititmhhienin s1 1a00y--3 3it0 0w y yileela atarrsks,e ,s s1oo0mm0e ey s esaayry si ti. t w Iwt iliilsl l tntaaokketeo 1 r1i0o00u0 sy yleyea arrss. .I It ti sis n nootoftoor riYoiououssrlkyly on a rainofyof S YYaootrurkkr d ooanny a ai s rr aamoiinnyyr e SS maofatatu uaprr dsdca haaayyn ildislsen kmomongoerwr.ee l eofofd aga e cc haahablollenuent grgoeea..d signs, driver behaviour, etc.