Model-based robot controllers require customized control-oriented models, involving expert knowledge and trial and error. Remarkably, the Koopman operator enables the control-oriented model identification through the input–output mapping set, breaking through the barriers of the customization services. However, in recent years, research on Koopman-based robot control has mostly focused on lifting function construction, deviating from the original intention of improving the controller performance. Thus, we propose a robot controller autogeneration framework using the Bayesian-based Koopman operator, significantly releasing labor and eliminating the design obstacle. First, we introduce the Koopman-based system identification method and offer the basic lifting function design criteria. Then, a Bayesian-based optimization strategy with resource allocation is designed, which allows for the simultaneous optimization of the lifting function and the controller. Next, taking model-predictive control (MPC) as an example, a mission-oriented controller autogeneration framework is developed. Simulation and experimental results indicate that, under various robots and data sources, the proposed framework can effectively generate the robot controllers and perform with a far greater level of mission accuracy than the unoptimized Koopman-based MPC. Meanwhile, the proposed technique exhibits an obvious compensation effect against disturbances, demonstrating its practicability in robot control.

It is critical for retail enterprises to select good sites or locations to open their stores, especially in current competitive retail market. However, evaluating the goodness of sites in real business applications is a complex problem. That is, how to judge whether the market around a store site is good? We don't know the exact mechanism of how a site can be good and it is hard to have correct site goodness values as supervised labels. The Retail Outlet Site Evaluation (ROSE) tool is designed to learn the site evaluation model by integrating city geographic & demographic data and two kinds of expert knowledge: sample preference and feature preference. The feature preference information can help greatly reduce the required number of sample preferences. It enables our application practicable because it is almost impossible to give such amount of sample preference pairs manually by experts when ranking hundreds of data points. In the experiment and case study part, we show that the ROSE tool can achieve good results and useful for users to do site evaluation work in real cases.

This paper is concerned with cooperative control of a kind of multiple fish-like microrobots. Most of previous work on multi-robot cooperation is focused on the terrestrial robots and seldom deals with underwater applications. In fact, the tasks in hydro-environment is more challenging than those in ground circumstances and need the cooperation of robots much more. In this paper, we investigate this problem in the framework of an adversarial game with several underwater microrobots. A fuzzy reinforcement learning approach is adopted to acquire cooperative behavior and a behavioral hierarchical architecture is proposed. We conduct extensive experiments to verify the effectiveness of the proposed algorithms

This paper is concerned with a novel problem of underwater cooperative transportation of multiple fishlike robots. A situation based action selection mechanism is proposed for the robots to transport a floating object to its destination. There are various control methods for cooperation of multi-robot system, but few of them deal with the underwater applications. In this paper, we first present the development of a series of biomimetic fish-like robots. Then, employing these robotic fish, we design and implement a cooperative transportation task utilizing a situation based action selection approach. We fully implement our method on three fish-like robots and make various experiments in a lab environment

This paper presents a new method for formation control of multiple underwater fish-like robots. Considering both geometrical and mechanical constrains of the fish-like robots and based on the leader-following approach, a curvature coordinate is introduced to describe relative positions between different members within formations. Both the static and dynamic formations are concerned. We conduct simulations and physical experiments to verify effectiveness of the proposed algorithms

This paper presents a new method for controlling a group of nonholonomic mobile robots to achieve predetermined formations without using global knowledge. Based on the dynamic leader-follower model, a reactive tracking controller is proposed to make each following robot maintain a desired pose to its leader, and the stability property of this controller is discussed using Lyapunov theory. By employing such controllers, the N-robot formation control problem can be decomposed into decentralized tracking problems between N-l followers and designated leaders. Additionally, graph theory is introduced to formalize general formation patterns in a simple but effective way and two types of switching between these formations are also proposed. Numerical simulations and physical robots experiments show the effectiveness of our approach.

This paper presents a hierarchical framework for controlling a group of biomimetic fish robots to achieve cooperative tasks. Based on our previous successful work on the design and development of a robotic fish prototype, we attempt further to investigate the cooperation in groups of these fish. Employing top-down design approach, we propose a hierarchical architecture consisting of five levels: task level, role (or mode) level, behavior level, action level and controller level, to formalize the processes from task decomposition, role assignments and control performance. Two typical cases are developed to demonstrate the feasibility of the architecture and corresponding experimental results show that high efficiency and much greater capabilities are exhibited when the fish try to cooperate.

This paper is concerned with the problem of motion planning for multiple robotic fish swimming together to achieve a disk-pushing task. Based on our previous successful work on the design and development of robotic fish prototypes, we attempt further to investigate how to implement cooperation in groups of these fish. The motivation of this work is that the capability of a single fish robot is always limited while there are many complex missions in practice which should be accomplished by effective cooperation of multiple fish. A disk-pushing task is employed as an example to demonstrate the design of cooperative behaviors as well as strategies. Corresponding experimental results show that higher efficiency and much greater capabilities can be obtained when these fish planning their motion properly.