CMG Method of Designing Self-Balancing Bicycles: a Review

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

CMG Method of Designing Self-Balancing Bicycles: A Review

Avinash Kumar1, Adeel beg2, Pulkit Jain3 1,2Pursuing bachelor degree program in mechatronics engineering in Chandigarh University,India 3Assistant Professor in Mechatronics engineering in Chandigarh University, India Email:[email protected], [email protected]

Abstract In recent years researchers are busy in concept of electric vehicles and the driverless cars. Now compared to the automated cars, autonomous bicycles are more affordable and environment friendly in nature which do not produce any harmful effect to the environment. The autonomous bicycles reduce human effort and are also considered best for the physically disabled person. This review paper explains the benefits of autonomous bicycles and its various methods of balancing. The biggest challenge in automating bicycle is the inherent problem of staying balanced. This paper gives knowledge about the modifications which can be carried out on conventional bicycles to make them autonomous. Various electrical components and their uses are also described in this paper. The design of the autonomous bicycle using Control Movement Gyroscope (CMG) to make it self-balance also described in this paper. The hardware and software development of autonomous bicycles are described in this paper.

Keywords: CMG; driverless bicycles; self-balance; Autonomous; balancing; development.

1 INTRODUCTION Self balancing and driving bicycles or autonomous bicycles becomes a most searched topic on internet in this year. It has many benefits on normal bicycles ,bicycles usually lead many injuries but it can reduce injuries. Generally everyone think that why not we try to make 4-wheeled or 3-wheeled autonomous vehicles rather than trying to make autonomous bicycles. So if we compare bicycles from other vehicles then there are some reasons. Bicycles does not use more track space like any other vehicles. It can be more flexible and easy to handle for rider as compared to any other vehicles. It can easily cut through obstacles as compared to 4-wheeled or 3-wheeled vehicle. It can be also useful for transportation in small areas like mines or research areas.Apart from the potential advantages bicycles are also a unique platform for research. For example, it can help us to understand the physical human- machine interaction. Because there are various points for connecting a bicycle to its rider(like handle bars, comfortable seat, and peddles).Due to these points we can attach various multiple sensors to track activities of rider.In India, recently held KPIT sparkle 2016(annual national design and development innovation contest) students of IIT Kharagpur who came with driverless bicycle , took the first prize. In this study we are going to discuss about design and prototypes to make bicycles self balance and driverless.

Fig1. A self balance bicycle[1]

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

1.1.Advantages: The Autonomous bicycles have a fast speed as compared to normal bicycles. So, this is easy for a fat and a beginner to get some significant speed going. These bicycles are easy to handle with a less effort or no effort , just move the steering where you want to go. It occupies less track space as compare to any other vehicles. It provides a comfortable ride to the riders .

Fig2. Beginner on a self-balance bicycle.[2]

It does not produce any type of environmental pollution. It helps in many small areas like research areas and mines as a transport. The blind person or a physically disable person can also drive these bicycles without any harm. It can also run on wireless phone network by using GPS technology.

Fig3. Flow chart of advantages.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

1.2. Disadvantages: Autonomous bicycles have more parts or components as compare to the normal bicycles. So, there is more chance of something going wrong. Weight of the autonomous bicycle is more as compare to normal bicycle. Expensive than a normal bicycle. Programming is done in this , so there is possibilities that something unexpected fault may happen. Many taxi drivers will loose their jobs because of lot of advantages of autonomous bicycle on taxi. So, there is decrease in demand of drivers. It would not be affordable by average person due to its price when it will newly introduce in market.

Fig4. Flow chart of disadvantages

2.REVIEW:

It contains methods to achieve self balance bicycles and how to make driverless bicycle and to understand the human-machine interactions in driverless bicycles. The goal of this section is to introduce related work with the help of these topics. 2.1.Methods to achieve self balance bicycles: Now,we are going to discuss about four methods to self -balancing bicycle based on the literature review and these four types are as follows: • Control moment gyroscope • Mass balancing • Steering control • Reaction wheel

• Control moment gyroscope: A CMG is an attitude control device often used in spacecraft attitude control[3]. A typical CMG Is made of one spinning rotor and one motorized gimbal that can change the rotor’s angular momentum, which can generate processive torque to balance the bicycle. Many researchers have utilized CMGs to balance bicycles[4,5]. According to gyroscopic theory it can be clear that by making a flywheel process, it generates a moment to oppose the gravitational moment and due to this our bicycle will balanced.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

Fig5. Parts of Control moment gyroscope.[6]

• Mass Balancing: Bicycles can also balance by using a massive balancer on the bicycle to achieve self-balance [7,8].The mass balancing can provide a small amount of torque besides the structure required for mass balancing is simpler as compared to CMG. Steering Control: Steering control is a method to achieve real time control of bicycle balance by adjusting the steering system in real time [9,10]. So the steering control will consume the less power as compared to the gyroscope , but this method sufficiently cannot provide balance at low forward velocities. Reaction Wheel: The reaction wheel cannot response fast and and the torque is limited, so it is only to balance a small type of bicycle.

Fig6. Reaction wheel balancing a small object.[11]

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

TABLE1. ADVANTAGE AND DISADVANTAGE OF DIFFERENT TECHNIQUES.

Sr.No Disadvantages Methods Advantages 1 It produces High energy consumption. CMG large torque to balance. 2 Mass It has Small torque. balancing simple structure as compared to CMG. 3 Steering It gives At low velocities control real time there is no response balance. and it consume less energy. 4 Reaction It Output torque is Wheel response limited. in short time period.

2.2. Driverless bicycles: In the recent year, the researchers show a huge interest on autonomous driving, mostly in autonomously driving cars. However, there are few researches that focus on autonomous driving bicycles. On the basis of pre-existing literature there are few projects on driverless bicycles are discussed below. Student of IIT Kharagpur develop a driverless bicycle that can reach you with just an SMS. It uses GPS for automatic respond to GPS coordinates of destination which are received through SMS with the help of an application which are meant for this bike. Laser and sonar-based sensor are used to avoid obstacles. It is enabled to follow a specialised bicycle lane due to a unique and affordable software architecture. It is connected to a wireless network which provides live tracking mechanism. A power-assisted electric bicycle was designed by researchers from University of California, Berkeley for uphill riding. Control algorithms in the system can provide ﬂexible assistive power through estimating and compensating for the changing environmental disturbances [12]. Automotive collision avoidance systems have been developed by many automotive manufacturers and a sensing system for bicyclist-motorist crash prediction is developed to `accurately track rear vehicles that can have two-dimensional motion [13]. In addition, vision-based motion planning systems for an autonomous motorcycle has been designed for desert terrain, where uniform road surface and lane markings are not present [14]. Finally, the study of self-driving and human-robot interaction can be developed by deep learning algorithm.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

3.HARDWARE MODIFICATION FOR AUTONOMOUS BICYCLE: The hardware for this project was created and designed to create a best self balance and the driverless bicycle. This section includes the modification of bicycle model, some useful sensors which is used to control the various actions of bicycles, steering control assembly, CMG development and various electrical components . This bicycle contains a hub motor of 300w power which allows the bicycles to carry large amount of weight of the components used in this model . A pre-built electric bicycle, the X-Treme Trail Maker Elite Mountain Bike, was selected based on several features. The overall design of the modiﬁcations to the original bicycle are illustrated in Figure 7.

Fig7. Prototype of self-balancing bike.

3.1. PARTS MODIFICATION: In order to implement a control algorithm some of the part of the bicycle is need to be modified and this can be done in following ways. The throttle signals of the bicycle are transferred from the handlebar to the microcontroller (it is an IC chips which is used to executes the program for automatically controlled products) and the servo is attached to rear brakes of bicycles which allows the bicycles to be braked without any physical contact

Fig8.Different types of microcontroller[15]

The hall effect sensors (used for proximity sensing, positioning, speed detection) are placed on the fork of each wheel of bicycles and three magnets should be stick to the spokes. Due to this microcontroller can easily measure the speed of bicycle.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

Fig9. Hall effect sensor.[16]

3.2. Control box and electrical components: A box is designed and 3d printed to give space to the electronics. The box is placed on the central triangle of the bicycle which allows a good control for all the components. On each corner of the box there is a hole which allows wires to be scattered easily. The cover of this box contains a 60mm fan which provides a cooling to all the components present inside the box. Dual buck converters: All the electronic components present in control box needs 5v to run. So, it is necessary to generate a 5v rail. These dual buck converters used to supply enough voltage. Raspberry Pi B: The image processing and high-level control of the bicycles is done on Raspberry Pi Model 3 B which was programmed in python. Custom user interface is used to start a command in the program and raspberry pi needs WIFI to communicate with custom user interface. After this we can see the real time information about the bicycles like speed, acceleration, location and direction.

Fig10.labelling of various components present in control box.

Phidgetspatial 3/3/3 IMU: It is an inertial measurement unit measures the acceleration (static and dynamic) in 3 axes, also measures magnetic field in the bicycle frame when it communicates through USB with raspberry Pi B.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

Adafruit ultimate GPS: For intensify the position of the bicycles the raspberry uses Adafruit GPS module over UART (A universal asynchronous receiver-transmitter which is used to transfer data between two devices) to get data. Psoc 5LP: it is a microcontroller which communicates with the raspberry over a serial connection for the low level controls of the bicycles. 3.3. Steering control assembly (RGB camera and Lidar): The steering motor is placed on the handle bars and there is encoder present in it . The gears is also placed at the handle under the housing. To control the steering in best way a modified steel gears with a pressure angle of 20 degree are attached to the motor shaft and the steering column just below the handle bars. It also includes a brace to mount the Lidar unit and raspberry camera extended on an armature. The lidar is present on the handle is used to detect the obstacles and the RGB camera is used to define the lines on the road. 3.4. CMG Development: Above the rear wheel of the bicycles ,a box containing the pair of control moment gyroscope(CMG). The CMG is used to provide the balance to the wheel at low forward speed. Two CMG will produce more torque as compare to the one CMG and they balance the bicycles easily. The gimbals will be moved by a sprocket and chain which will be operate by using a high torque servo. This assembly contains a 60mm fan in front of box and 140mm exhaust fan at rear of the box which is used to cool the motors and microcontrollers present in CMG. Each of CMG gimbal contains flywheel having weight 6.8kg which will rotate at 8000rpm which will produce torque for balancing force.

Fig11. Model of CMG assembly.

Fig12. Block diagram for the smart bicycles.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

4. SOFTWARE DEVELOPMENT:

In this section the software system is divided into four main subsystems which are as follows high level control, actuator control, human sensing and remote user interface. This software system is implemented entirely on low-cost commercial hardware, using the C and Python programming languages. The high level control includes raspberry Pi 3B which are responsible for balancing and navigate the bicycles. The input to this control system contains data from LIDAR, IMU and RGB camera as shown in fig10. The data which are obtained from this system are use to find the direction of bicycles and control the movement of the bicycles. The actuator control includes PSoc 5LP microcontrollers. This microcontrollers receive the input for the angle of steering, speed of bicycle and braking amount over a serial connection to the raspberry pi and adjust actuator to get these values. The actuator psoc 5lp also send the data back to the Raspberry pi 3b about the original condition of the bicycles through high level control algorithms as shown in fig10. The remote user interface subsystem contains two primary components: an RC transmitter and a tablet or computer. During the operation RC transmitter is used to provide the manual input to the bicycles via 2.4GHZ RC receiver which are connected to the Psoc 5LP through Futaba S.Bus protocol. If any dangerous situation will arise during testing the actuators on the bicycles (throttle and break) can be actuate remotely . The raspberry pi connected to the mobile device through wireless network to display real time data.

CONCLUSION: Balancing of two wheeled vehicles plays a major role in the complex transportation system across the World. In this paper a design for self-balancing bicycle was presented and discussed. In this paper we focused on the various conditions to stabilizing the balance of the bicycle like maintaining the centre of gravity, and by controlling the falling angles direction. The primary goals of this study to make a hardware modification of the autonomous bicycles and define the electrical components with different applications used in it and to build a research platform for the further study of autonomous bicycles, algorithms to balance, and human-machine interaction. In this attempt, pair of control moment gyroscope(CMG) is placed on the rear wheel of the bicycle to get extra torque for balancing. These CMG is providing the balancing actuation for the bicycles at low forward speed. The raspberry pi 3b are used as a central control and Psoc 5LP are used as a actuator control in this method. The raspberry pi 3b are used for balancing and navigating the bicycles. It takes the input from the components through asynchronous process and defines the path of bicycles. The actuator control is used to manage the bicycle speed and steering angle. This study contains potential to generate data which can be used in both self-driving algorithms and human robot-interactions.

REFERENCES

[1]self balance bicycle. Available online: https://www.youtube.com/watch?v=4UHLRmYNDMQ/ accesed on 11-11-2019. [2] https://www.livemint.com/Leisure/l9OYXq6qW38sdehmLUrpFJ/Jyrobike-The-bike-that-never- tumbles.html [3] Roser, X., and Sghedoni, M., 1997. “Control moment gyroscopes (cmg’s) and their application in future scientiﬁc missions”. In Spacecraft Guidance, Navigation and Control Systems, Vol. 381, p. 523. [4] Yetkin, H., Kalouche, S., Vernier, M., Colvin, G., Redmill, K., and Ozguner, U., 2014. “Gyroscopic stabilization of an unmanned bicycle”. In American Control Conference (ACC), 2014, IEEE, pp. 4549–4554. [5] Beznos, A., Formal’Sky, A., Gurﬁnkel, E., Jicharev, D., Lensky, A., Savitsky, K., and Tchesalin, L., 1998. “Control of autonomous motion of two-wheel bicycle with gyroscopic stabilisation”. In Robotics and Automation, 1998. Proceedings. 1998 IEEE International Conference on, Vol. 3, IEEE, pp. 2670–2675. [6].CMG. available online: https://imgbin.com/png/7y3zSam9/gyroscope-inertia-gimbal-lock-rotation-png.

International Journal of Advance Science and Technology Vol. 29, No. 10S, (2020), pp.5928 -5937

[7] Lee, S., and Ham, W., 2002. “Self stabilizing strategy in tracking control of unmanned electric bicycle with mass balance”. In Intelligent Robots and Systems, 2002. IEEE/RSJ International Conference on, Vol. 3, IEEE, pp. 2200–2205. [8] Yamakita, M., and Utano, A., 2005. “Automatic control of bicycles with a balancer”. In Advanced Intelligent Mechatronics.Proceedings,2005IEEE/ASME International Conference on, IEEE, pp. 1245–1250. [9] Getz, N. H., and Marsden, J. E., 1995. “Control for an autonomous bicycle”. In Robotics and Automation, 1995. Proceedings., 1995 IEEE International Conference on, Vol. 2, IEEE, pp. 1397–1402. [10] Tanaka, Y., and Murakami, T., 2004. “Self-sustaining bicycle robot with steering controller”. In Advanced Motion Control,2004.AMC’04.The8thIEEEInternationalWorkshop on, IEEE, pp. 193–197. [11] Reaction wheel balance a small object. Available onine: https://robohub.org/swiss-robots-cubli-a-cube-that- can-jump-up-balance-and-walk-across-your-desk/ [12] Fan, X., and Tomizuka, M., 2010. “Robust disturbance observer design for a power-assist electric bicycle”. In American Control Conference (ACC), 2010, IEEE, pp. 1166– 1171. [13] Jeon, W., and Rajamani, R., 2017. “Two-dimensional active sensing system for bicyclist-motorist crash prediction”. In American Control Conference (ACC), 2017, IEEE, pp. 2315–2320. [14] Song, D., Lee, H. N., Yi, J., and Levandowski, A., 2007. “Vision-based motion planning for an autonomous motorcycle on ill-structured roads”. Autonomous Robots, 23(3), pp. 197–212. [15]https://www.google.com/search?q=microcontroller&rlz=1C1CHBF_enIN848IN848&sxsrf=ACYBGNQw 0LswMKuh595xdkZ3Hs5cE10mSQ:1573371201047&source=lnms&tbm=isch&sa=X&ved=0ahUKEwiz1aa 8kN_lAhXXbSsKHXc6AvUQ_AUIEigB&biw=1396&bih=686#imgdii=HsEFRXqI4ECATM:&imgrc=HsEF RXqI4ECATM: [16]https://www.google.com/search?q=hall+effect+sensors&sxsrf=ACYBGNTra1JUALVB_uA7_n_ttbYEjwj xDQ:1573371659200&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjGkeKWkt_lAhUDaI8KHczzC1QQ _AUIEygC&biw=1396&bih=686#imgdii=0HaCRb8qsmNeeM:&imgrc=jOwhY_DEc8MviM: