Human guided cooperative robotic agents in smart home using beetle antennae search  

作　　者：Ameer Tamoor KHAN Shuai LI Xinwei CAO 

机构地区：[1]Department of Computing,The Hong Kong Polytechnic University,Hong Kong 999077,China [2]School of Information Science and Engineering,Lanzhou University,Lanzhou 730000,China [3]School of Management,Shanghai University,Shanghai 201900,China

出　　处：《Science China(Information Sciences)》2022年第2期85-101,共17页中国科学（信息科学）（英文版）

基　　金：partially supported by CAAIXSJLJJ-2020-012A。

摘　　要：In this paper,we propose a control framework for cooperative robotic agents,which constitutes an essential component in the construction of futuristic smart-homes.Such agents assist humans in efficiently completing household chores.Usability,human friendliness,autonomy,and intelligent decisionmaking are the top considerations for designing such a system along with reliability,accuracy,and efficiency.Implementing a distributed control algorithm considering these goals is a complicated task since classical control frameworks focus on specialized robots working in an industrial environment and do not capture unique features of the home environment.For example,a household robotic agent needs to perform several general-purpose tasks without assuming that the user has specialized training similar to an industrial operator.Since the challenges and goals in designing a household robotic agent are different,there is a need for a control framework centered around the required goals.The proposed control framework considers the collision problem between several cooperative robotic agents while assisting the human user.We propose an optimization-driven approach to avoid static and dynamic obstacles present in the environment while simultaneously controlling the robots as commanded by the user.We formulate the optimization problem that incorporates the required goals and then use a neural network to solve the optimization problem efficiently.The neural network,beetle antennae search zeroing neural network(BASZNN),is inspired by the natural behavior of beetles.It solves the optimization problem in a gradient-free manner contributing to the computational efficiency of the neural network.Additionally,the distributed-processing capability of the neural networks contributes to computational efficiency and matches the distributed nature of the underlying problem.For testing the performance of BASZNN,we use V-REP and MATLAB to simulate a household environment.Three cooperative agents(KUKA LBR IIWA 7 mounted on P3-DX)assist a person i

关 键 词：smart-home assistive agents metaheuristic optimization beetle antennae search zeroing neural network human guided 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TU855[自动化与计算机技术—控制科学与工程] TP242[建筑科学]