UNDERWATER DRONES CONTROL TOWER UNDERWATER DRONES CONTROL TOWER Adnan Tahirovic - Kemal Delic
ARCHITECTURE, DESIGN, ENGINEERING Talk outline - why and how uwr is important ?
+ Underwater World Explained + Technology Challenges and Future Developments + Architecture, Design, Engineering Introduced + Control Tower Architecture : Conceptual View + Cloud Based Analytics : Design View + Mapping the Sea Bottom - Navigation + Intelligent Underwater Drone Design : Swarming + Live Demo : Navigation + Algorithm Explained + Future Explored Underwater World Explained
Hidden Face of Oceans Oceans in numbers
Depth of Ignorance - Level of Dependency LIVING SPECIES
91% still unknown, 13% only catalogued
RESOURCES
90% of transportation , 20% of animal proteins
PHYSICAL WORLD
71% Earth surface water, 93% of heat stored
80% of volcanic activities under water
Eight challenges - unknowns
+ Thermal - temperature anomalies, impacting life on earth + Geological - cobalt, platinum etc found in proximity of volcanoes + Genetic - new molecules for drugs learned from deep sea creatures + Ecological - destruction of habitats + Climate - likely impact of human activities + Hydrodynamic - Golf Stream 15% lower circulation + Chemical - seas and oceans may become toxic or sterile + Physical - rise of the sea level, 60% megapolis on the coast Map of unknown worlds
Nobody knows for sure what might be below seabed where, how much it is worth
Strategic future exploration
Always preceded by map creation New molecules
Discoveries of new drugs and materials
Species living in extreme conditions, no sunlight, huge pressure and cold
Gulf stream change
Climate Change
Might be caused by
Changes in ocean
Streams ?
Last 3 centuries Underwater World stratified
Into the abyss
0 m 4000 m ALVIN (USA) 8145 m deepest fish (snailfish) found 100 m 8400 m Puerto Rico trench Depths for divers Sentry HROV (USA) (deepest point in the Atlantic Ocean) 500 m MIR (Russia) Nautile (France)
1000 m Depths for submarines 6000 m 97% of ocean depths are less than 6000m ABISMO ROV (Japan) 2000 m Shinkai 6500 (Japan) 6500 m 10350 m Jason ROV (USA)
11000 m Mariana Trench 3000 m Maximal dive for whales 7000 m Jiaolong (China) (Deepest point in the Pacific Ocean) (Curver beaked whale) Mariana Trench reached by: Bathyscaph Trieste on 23.1.1960. Kaiko ROV on 24.3.1995. lost in 2003 2 Nereus HROV on 31.5.2009. lost in 2012 Area: 361,000,000 km (71% of Earth surface) Deepsea challenger on 26.3.2012. Technology Challenges and Future Developments
Last week in Croatia ..
Breaking The Surface 2018
10 Years anniversary workshop Game changing technologies that have the potential to significantly enhance capabilities of systems and transform how we will use future systems
✓ Quantum computing, neuromorphic (brain inspired computing), … ✓ Microelectronics, (components built of molecules), … ✓ Robotics, Soft reconfigurable robotics, …. ✓ Nanomaterials, advanced materials,…. ✓ Genetics, …. ✓ “Big data” , … ✓ Alternative energy sources, ✓ Artificial intelligence, machine learning, …. ✓ Modeling and simulation, .... Ec subucultron project http://www.subcultron.eu/
Venice - Laguna
Health Monitoring
Architecture, Design, Engineering Introduced
Control Tower Architecture : Conceptual View
Control Tower Architecture : Conceptual View
CT: Cloud Based Analytics : Design View
Intelligent Underwater Drone Design : concept
Vision dl explained
Picture
Recognition Olfaction with DL NN
Odor Recognition Intelligent Underwater Drone Design
Deep dive follows ..
A possible mission: Coverage path planning (CPp)
• Monitoring • Surveillance • Hazard detection • Planetary exploration • Rescue • Cleaning • De-mining • Fire extinguishing • Agricultural spraying
Tahirovic Adnan and Alessandro Astolfi. "A convergent solution to the multi-vehicle coverage problem." American Control Conference (ACC), 2013. IEEE, 2013. CPp Algorithm Explained
• Unconstrained environment • Fully-connected swarm CPp Algorithm Explained CPp Algorithm Explained Partially-connected swarm CPp Algorithm Explained
Tahirovic Adnan, et al. "A receding horizon scheme for constrained multi-vehicle coverage problems." Systems, Man, and Cybernetics (SMC), 2016 IEEE International Conference on. IEEE, 2016. CPp Algorithm Simulation CPp Algorithm Features
• Simple!!! • Cooperative!!! • Scalable!!! • Robust!!! • Adaptive???
All necessary features of swarm intelligence obtained via simple agent’s rules. Rapidly Exploring Random Trees Rapidly exploring random vines
Tahirovic Adnan, and Mina Ferizbegovic. "Rapidly-Exploring Random Vines (RRV) for Motion Planning in Configuration Spaces with Narrow Passages." 2018 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2018. Rapidly exploring random vines Rapidly exploring random vines Rapidly exploring random vines Rapidly exploring random vines Key takeaways
+ Shift from ROV/AUV devices to the entire ecosystem + The rise of multi-modal systems : flying, sailing, diving + Big Data collections waiting for better analytics + Security nearly non-existent + Biology inspired sensory and communication systems
+ Ocean explorations will be even more important in the future and AI approaches and ML methods will play crucial role - from intelligent swarm edge devices to elaborate analytics in the cloud Back up slides Future explored - what’s next? • Monitoring of Bosnian lakes • Health monitoring of Venetian Lagoon • Monitoring of fishponds in Norway • Monitoring underwater cables