Artificial Intelligence in Wireless Robotics

320 Index Position estimation 189, 209 Position tracking 189, 199 O Posteriori probability 62, 94, 169, Object-level space 47 Observation model 181, 202, 203, 221 Potential functions 222 204 p-persistent 310, 311, 312 Observations 60, 175, 202, 255 Prediction 13, 156, 184, 264 One-shot decision 81, 95, 278 Principal component analysis Open system interconnection 292 Optical flow 250 (PCA) 67, 70 Optimal assignment (OA) 278 Priori probability 7, 36, 90, 94 Optimal planning 227, 228 Priority queue 32, 35, 44 Optimal solution 26, 47, 96, 279 Private reference 189, 235, 309, 311 Optimal state-value function 100, Probabilistic graphical models 218, 137 223 Optimally efficient 46 Probabilistic robotics 174 Optimization 18, 51, 109, 311 Probability mass function Order 19, 67, 223, 286 Orthogonal principle 166, 182 (PMF) 106 Outliers 67, 76, 77 Problem 5, 45, 115, 281 Projection theorem 166 P Proprioceptive sensors 12 Parent node 29, 31 Public reference 189, 233, 306, 307 Partially observable 10, 138 Partially observed MDP Q Q-learning 138, 159, 233, 298 (POMDP) 139 Quadratic loss 7 Passive localization 190 Quadrature phase shift keying 290 Path 19, 47, 141, 165 Queue 32, 44, 235, 283 Path cost 26, 28, 35, 45 Path planning 228, 229 R Pattern Recognition 63, 73, 85, 89 Radio detection and ranging Performance measure 7, 17, 269, (RADAR) 92, 248 288 Radio Resource Units (RRUs) 301 Phase estimation 291 Ranking 85 Phase interferometry 194 Rational agent 6, 7, 14 Phase shift keying 290 Real-time ALOHA 302, 310, 312 Planning agents 25 Received signal strength (RSS) 196 Policy 10, 103, 158, 297 Receiver 2, 86, 197, 308 Policy evaluation 108 Receiver operating curve (ROC) 93 Policy improvement 108 Recently-weighted average 132 Policy iteration 101, 137

Recurrent neural network (RNN) 75 Index 321 Recursive best-first search Size 5, 54, 237, 273 (RBFS) 47 Skewness 78 Reducing nodes (R-nodes) 261 Smart factory 14, 283, 286, 287 Regression 51, 59, 64, 273 Smart manufacturing 2, 5, 283, 312 Regret 115, 116 Software agent 6, 12 Regularization 51, 67 Solution 4, 30, 69, 285 Reinforcement learning 96, 136, Source coding 289 Spanning tree 41 161, 263 Squashing function field 73 Relation 15, 112, 197, 249 Stable 8, 69, 304, 305 Relaxed problem 47 Stack 32, 207, 217 Repeated state 30, 31 State 6, 28, 135, 298 Residual sum of squares (RSS) 56, State machine 9 State space 7, 47, 203, 262 167 State transition diagram 9, 18, 104, Reward 96, 135, 232, 308 Ridge regression 167, 168 125 Right-skewed 78 State transition probability 101, Root 19, 42, 261, 278 124, 174, 231 S State-transition probabilities 124 Scalability 293, 300 State-value function 98, 135, 158, Search 5, 33, 107, 261 Search Tree 29, 32, 42, 47 308 Semantic graphs 224, 225 Static environments 190, 199 Sensing nodes (S-nodes) 261 Stationary 3, 132, 165, 204 Sensor 6, 177, 192, 270 Statistical decision theory 7, 81, 85, Sensor model 13 Sensor network localization 189 246 Sequential decision 81, 96, 132, 252 Statistical inference 79, 85, 163, Sequential detection 94 Shortest path 38, 108, 120, 229 168 Signal space 88, 89, 91, 290 Statistical learning 51, 54, 175 Signal-to-noise ratio (SNR) 84, 291 Stereo camera 205 Simplified MA* (SMA*) 47 Stereo vision 250 Simultaneous localization and Stopping rule 94, 95, 259 Stopping time 94 mappingb (SLAM) 200 Strategy 33, 106, 252, 269 Single-frame 250 Structural risk minimization Single-robot 278 Single-task 278 (SRM) 54 Successor 26, 45, 100, 140 Sum of squares regression (SSR) 60 Sum of squares total (SST) 59

322 Index Univariate tree 256 Unmanned aerial vehicle (UAV) Supervised learning 55, 67, 255, 272 215 Unobservable 10 Support vector machines (SVMs) Upper confidence bounds (UCB) 54 117 Synapses 73 Utility function 7, 85, 280 Synchronization 84, 289, 291 System Dynamics 8, 9, 103, 252 V Value iteration 101, 137, 149, 264 T Variable 14, 50, 78, 223 Target area 230, 231 Vehicle-to-infrastructure (V2I) 147, Target policy 151 Temporal relationship 225 298, 299, 300 Temporal-difference (TD) Vehicle-to-infrastructure-to- learning 150 vehicle (V2I2V) 298, 300, 303, Terminal rewards 98 304 Terminal state 104, 150, 155, 159 Vehicle-to-vehicle (V2V) 298, 300, Thompson sampling 118, 120, 126 303, 304 Time delay estimation (TDE) 196 Verification tour 260 Time-difference-of-arrivals Vertex 19, 119, 218, 227 Vertical federated learning 270 (TDOA) 196 Viterbi algorithm 43, 176 Time-extended assignment 278 Time-of-arrival (TOA) 190 W Timing recovery 197, 291 Weighted least-squares 51 Total assignment 14, 18 Wiener filter 3, 58, 179 Tractability 5 Wiener-Hopf Equation 58, 167 Tragedy of commons 307 WiFi 296 Transition model 13, 26, 29, 202 Wireless local area networks 106, Transmitter 83, 195, 289, 295 Tree 19, 46, 147, 303 296 Trellis coded modulations 43 Wireless robotics 252, 275, Truncated quadratic loss 7 Turing machine 3, 5 288, 293 Wireless sensor networks U Unbiased 169, 170, 172 (WSNs) 172, 209, 275 Undirected graphical model 222 World 3, 78, 163, 287 Uniform search 33, 46 Uniform-cost search 35, 45, 46, 47 Z Z-Score Standardization 78

About the Author Kwang-Cheng Chen has been a Professor at the Department of Electrical Engineering, University of South Florida, since 2016. From 1987 to 2016, Dr. Chen worked with SSE, Communications Satellite Corp., IBM Thomas J. Watson Research Center, National Tsing Hua University, HP Labs., and National Taiwan University in mobile communications and networks. He vis- ited TU Delft (1998), Aalborg University (2008), Sungkyunkwan University (2013), and Massachusetts Institute of Technology (2012–2013, 2015–2016). He founded a wireless IC design company in 2001, which was acquired by MediaTek Inc. He has been actively involving in the organization of various IEEE conferences and serving editorships with a few IEEE journals (most recently as a series editor on Data Science and AI for Communications in the IEEE Communications Magazine), together with various IEEE volunteer ser- vices to the IEEE, Communications Society, Vehicular Technology Society, and Signal Processing Society, such as founding the Technical Committee on Social Networks in the IEEE Communications Society. Dr. Chen also has contributed essential technology to various international standards, namely IEEE 802 wireless LANs, Bluetooth, LTE and LTE-A, 5G-NR, and ITU-T FG ML5G. He has authored and co-authored over 300 IEEE publications, 4 books published by Wiley and River, and more than 24 granted US patents. Dr. Chen is an IEEE Fellow and has received a number of awards including 2011 IEEE COMSOC WTC Recognition Award, 2014 IEEE Jack Neubauer Memorial Award, 2014 IEEE COMSOC AP Outstanding Paper Award. Dr. Chen’s current research interests include wireless networks, arti- ficial intelligence and machine learning, IoT/CPS, social networks and data analytics, and cybersecurity. 323



ARTIFICIAL INTELLIGENCE IN WIRELESS ROBOTICS Kwang-Cheng Chen Robots, autonomous vehicles, unmanned aerial vehicles, and smart factory, will significantly change human living style in digital society. Artificial Intelligence in Wireless Robotics introduces how wireless communications and networking technology enhances facilitation of artificial intelligence in robotics, which bridges basic multi-disciplinary knowledge among artificial intelligence, wireless communications, computing, and control in robotics. A unique aspect of the book is to introduce applying communication and signal processing techniques to enhance traditional artificial intelligence in robotics and multi-agent systems. The technical contents of this book include fundamental knowledge in robotics, cyber-physical systems, artificial intelligence, statistical decision and Markov decision process, reinforcement learning, state estimation, localization, computer vision and multi-modal data fusion, robot planning, multi-agent systems, networked multi-agent systems, security and robustness of networked robots, and ultra-reliable and low-latency machine-to-machine networking. Examples and exercises are provided for easy and effective comprehension. Engineers wishing to extend knowledge in the robotics, AI, and wireless communications, would be benefited from this book. In the meantime, the book is ready as a textbook for senior undergraduate students or first-year graduate students in electrical engineering, computer engineering, computer science, and general engineering students. The readers of this book shall have basic knowledge in undergraduate probability and linear algebra, and basic programming capability, in order to enjoy deep reading. River Publishers