Optimizing a prototype of a reconfigurable mobile robot using analytical geometry projections and particle swarm optimization

To study non-linear relationships between reconfigurable variables governing prototypes of extensively reconfigurable wheeled mobile robots, a model describing the main stability metrics of the mobile robots was established by combining an analytical method and a geometrical projection method. An optimized reconfigurable robot prototype was obtained using particle swarm optimization (PSO) on the premise of a maximum equivalent stability margin. The optimization results showed better consistency than those of reconfiguration tests of mobile robots. Reconfigurable prototypes of mobile robots should comply with geometric symmetry and the balanced load principle in order to allow reconfigured robots to progress unaided in a straight line. Reconfigurable parameters should create a margin of safety so as to keep the robot working well. Within this margin, the principle of trade-offs should be applied to various parameters. Extensively reconfigurable mobile robots with non-linear characteristics of stability are of great theoretical significance and have applied value in performance prediction-a key requirement for task planning and motion control of robots.