This is the latest version of this item.

Abstract

Recently, cooperative control of multiple unmanned vehicles has attracted a great deal of attention from scientific, industrial, and military aspects. Groups of unmanned ground, aerial, or marine vehicles working cooperatively lead to many advantages in a variety of applications such as: surveillance, search and exploration, cooperative reconnaissance, environmental monitoring, and cooperative manipulation, respectively. During mission execution, unmanned systems should travel autonomously between different locations, maintain a pre-defined formation shape, avoid collisions of obstacles and also other team members, and accommodate occurred faults and mitigate their negative effect on mission execution.
The main objectives of this dissertation are to design novel algorithms for single wheeled mobile robots (WMRs) trajectory tracking, cooperative control and obstacle avoidance of WMRs in fault-free situations. In addition, novel algorithms are developed for fault-tolerant cooperative control (FTCC) with integration of fault detection and diagnosis (FDD) scheme.
In normal/fault-free cases, an integrated approach combining input-output feedback linearization and distributed model predictive control (MPC) techniques is designed and implemented on a team of WMRs to accomplish the trajectory tracking as well as the cooperative task. An obstacle avoidance algorithm based on mechanical impedance principle is proposed to avoid potential collisions of surrounding obstacles. Moreover, the proposed control algorithm is implemented to a team of WMRs for pairing with a team of unmanned aerial vehicles (UAVs) for forest monitoring and fire detection applications.
When actuator faults occur in one of the robots, two cases are explicitly considered: i) if the faulty robot cannot complete its assigned task due to a severe fault, then the faulty robot has to get out from the formation mission, and an FTCC strategy is designed such that the tasks of the WMRs team are re-assigned to the remaining healthy robots to complete the mission with graceful performance degradation. Two methods are used to investigate this case: the Graph Theory, and formulating the FTCC problem as an optimal assignment problem; and ii) if the faulty robot can continue the mission with degraded performance, then the other team members reconfigure the controllers considering the capability of the faulty robot. Thus, the FTCC strategy is designed to re-coordinate the motion of each robot in the team. Within the proposed scheme, an FDD unit using a two-stage Kalman filter (TSKF) to detect and diagnose actuator faults is presented.
In case of using any other nonlinear controller in fault-free case rather than MPC, and in case of severe fault occurrence, another FTCC strategy is presented. First, the new reconfiguration is formulated by an optimal assignment problem where each healthy WMR is assigned to a unique place. Second, the new formation can be reconfigured, while the objective is to minimize the time to achieve the new formation within the constraints of the WMRs' dynamics and collision avoidance. A hybrid approach of control parametrization and time discretization (CPTD) and particle swarm optimization (PSO) is proposed to address this problem. Since PSO cannot solve the continuous control inputs, CPTD is adopted to provide an approximate piecewise linearization of the control inputs. Therefore, PSO can be adopted to find the global optimum solution.
In all cases, formation operation of the robot team is based on a leader-follower approach, whilst the control algorithm is implemented in a distributed manner. The results of the numerical simulations and real experiments demonstrate the effectiveness of the proposed algorithms in various scenarios.

Available Versions of this Item

• Cooperative Control of Multiple Wheeled Mobile Robots: Normal and Faulty Situations. (deposited 27 Oct 2022 13:50)
  • Cooperative Control of Multiple Wheeled Mobile Robots: Normal and Faulty Situations. (deposited 19 Nov 2019 16:37) [Currently Displayed]