A 65-nm 8-to-3-b 1.0–0.36-V 9.1–1.1-TOPS/W Hybrid-Digital-Mixed-Signal Computing Platform for Accelerating Swarm Robotics

Low-power edge-intelligence is leading to spectacular advances in smart sensors, actuators, and human–machine interfaces. In particular, energy efficiency is driving key advances in robotics, where low-power computation is augmented with smart control and mechanical systems to enable small-sized and intelligent drones, unmanned aerial vehicles (UAVs), micro-sized cars, and so on with applications in surveillance, disaster relief, and reconnaissance. Furthermore, for a variety of tasks, swarms of robots are often used as opposed to the individual robots. This article presents an energy-efficient computing platform that can enable a sample class of algorithms for swarm robotics. We demonstrate that both physical-model-based algorithms as well as learning-based algorithms can be supported on the same computing platform. We also demonstrate that with changing swarm sizes, the number of bits required to compute also scales. We take advantage of this observation to propose a hybrid-digital-mixed-signal computing platform, whose energy efficiency scales with the resolution of the data path and hence the swarm size. Measurements on a 65-nm CMOS test-chip demonstrate a peak energy efficiency of 9.1 TOPS/W at a 3-b resolution, and it scales down to 1.1 TOPS/W at an 8-b resolution.

[1]  Charles W. Warren,et al.  Multiple robot path coordination using artificial potential fields , 1990, Proceedings., IEEE International Conference on Robotics and Automation.

[2]  Kay Chen Tan,et al.  Evolutionary artificial potential fields and their application in real time robot path planning , 2000, Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512).

[3]  Qidan Zhu,et al.  Robot Path Planning Based on Artificial Potential Field Approach with Simulated Annealing , 2006, Sixth International Conference on Intelligent Systems Design and Applications.

[4]  Arijit Raychowdhury,et al.  14.1 A 65nm 1.1-to-9.1TOPS/W Hybrid-Digital-Mixed-Signal Computing Platform for Accelerating Model-Based and Model-Free Swarm Robotics , 2019, 2019 IEEE International Solid- State Circuits Conference - (ISSCC).

[5]  Marian Verhelst,et al.  14.5 Envision: A 0.26-to-10TOPS/W subword-parallel dynamic-voltage-accuracy-frequency-scalable Convolutional Neural Network processor in 28nm FDSOI , 2017, 2017 IEEE International Solid-State Circuits Conference (ISSCC).

[6]  Hiroyuki Kobayashi,et al.  An LDPC Decoder With Time-Domain Analog and Digital Mixed-Signal Processing , 2014, IEEE Journal of Solid-State Circuits.

[7]  Sergey Levine,et al.  Self-Supervised Deep Reinforcement Learning with Generalized Computation Graphs for Robot Navigation , 2017, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[8]  S. Peng.  A Generalized dynamic programming principle and hamilton-jacobi-bellman equation , 1992 .

[9]  Gu-Yeon Wei,et al.  DNN Engine: A 28-nm Timing-Error Tolerant Sparse Deep Neural Network Processor for IoT Applications , 2018, IEEE Journal of Solid-State Circuits.

[10]  Minoru Asada,et al.  Purposive Behavior Acquisition for a Real Robot by Vision-Based Reinforcement Learning , 2005, Machine Learning.

[11]  Arijit Raychowdhury,et al.  A 55nm time-domain mixed-signal neuromorphic accelerator with stochastic synapses and embedded reinforcement learning for autonomous micro-robots , 2018, 2018 IEEE International Solid - State Circuits Conference - (ISSCC).

[12]  Weihua Sheng,et al.  Multirobot Cooperative Learning for Predator Avoidance , 2015, IEEE Transactions on Control Systems Technology.

[13]  D.P. Stormont.  Autonomous rescue robot swarms for first responders , 2005, CIHSPS 2005. Proceedings of the 2005 IEEE International Conference on Computational Intelligence for Homeland Security and Personal Safety, 2005..

[14]  Gu-Yeon Wei,et al.  14.3 A 28nm SoC with a 1.2GHz 568nJ/prediction sparse deep-neural-network engine with >0.1 timing error rate tolerance for IoT applications , 2017, 2017 IEEE International Solid-State Circuits Conference (ISSCC).

[15]  Jaydeep P. Kulkarni,et al.  14.4 All-Digital Time-Domain CNN Engine Using Bidirectional Memory Delay Lines for Energy-Efficient Edge Computing , 2019, 2019 IEEE International Solid- State Circuits Conference - (ISSCC).

[16]  Bart De Schutter,et al.  A Comprehensive Survey of Multiagent Reinforcement Learning , 2008, IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews).

[17]  Marco Dorigo,et al.  Revision History , 2003 .

[18]  Jie Gu,et al.  19.7 A Scalable Pipelined Time-Domain DTW Engine for Time-Series Classification Using Multibit Time Flip-Flops With 140Giga-Cell-Updates/s Throughput , 2019, 2019 IEEE International Solid- State Circuits Conference - (ISSCC).

[19]  Arijit Raychowdhury,et al.  NavREn-Rl: Learning to fly in real environment via end-to-end deep reinforcement learning using monocular images , 2018, 2018 25th International Conference on Mechatronics and Machine Vision in Practice (M2VIP).

[20]  Sergey Levine,et al.  Collective robot reinforcement learning with distributed asynchronous guided policy search , 2016, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[21]  Alcherio Martinoli,et al.  Inspiring and Modeling Multi-Robot Search with Particle Swarm Optimization , 2007, 2007 IEEE Swarm Intelligence Symposium.

[22]  Jean-Claude Latombe,et al.  Numerical potential field techniques for robot path planning , 1991, Fifth International Conference on Advanced Robotics 'Robots in Unstructured Environments.

[23]  Min Cheol Lee,et al.  Artificial potential field based path planning for mobile robots using a virtual obstacle concept , 2003, Proceedings 2003 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM 2003).

[24]  Arijit Raychowdhury,et al.  A 55-nm, 1.0–0.4V, 1.25-pJ/MAC Time-Domain Mixed-Signal Neuromorphic Accelerator With Stochastic Synapses for Reinforcement Learning in Autonomous Mobile Robots , 2019, IEEE Journal of Solid-State Circuits.

[25]  Lei Liu,et al.  A Survey of Swarm Robotics System , 2012, ICSI.

[26]  Spring Berman,et al.  Optimized Stochastic Policies for Task Allocation in Swarms of Robots , 2009, IEEE Transactions on Robotics.

[27]  Jun-Seok Park,et al.  14.6 A 1.42TOPS/W deep convolutional neural network recognition processor for intelligent IoE systems , 2016, 2016 IEEE International Solid-State Circuits Conference (ISSCC).