| The smart wheelchair with autonomous functions can effectively enhance the safety and comfortableness. The smart wheelchair prototype is a typical application of the intelligent service robot system, which is developed for handicapped and the elderly people.A smart wheelchair is developed for the handicapped and the elderly. Based on powered wheelchair, the smart wheelchair equipped with a multi-sensors system including sonar, IR sensors, bumpers and vision, as well as an embedded control system for navigation and obstacle avoidance. The smart wheelchair can work in manual mode, semi-autonomous mode, PC mode and autonomous mode. It owns obstacles avoidance, collisions avoidance, floor unevenness detection, goal tracking and other functions.The TI DSP TMS320LF2407A is employed as the kernel of the embedded controller of the smart wheelchair. The hardware design includes motion control, the sensors’drivers and UART communication. It owns A/D sampling, D/A conversion, sixteen GPI/O control, the communication bus with Sonar Driver and other functions. The hardware design lays a solid foundation for the whole system development. All the programs of embedded software are developed by C language in the TI DSP integrated development environment CCS (Code Composer Studio). The architecture of the embedded software is a big loop with cascaded interrupts, which can quickly response many events happened in real-time. In software design, all the hardware resources are fully utilized. The software’s practicality and efficiency are carefully considered. The whole system has a stable performance and good expansibility.The smart wheelchair’s embedded control system follows the behavior-based programming idea. It uses Motor Schema thought to design the smart wheelchair’s basic behaviors. It adopts FSM (Finite State Machine) to describe the smart wheelchair’s tasks. By this way, the design facilitates the coordination of high-level planning and low-level real-time control. So that it makes the smart wheelchair not only owns high intelligence but also has fast reaction capability.Several field tests and experiments results analyses are carefully designed for the verification of smart wheelchair’s key performance. The results of distance sensors test and obstacle avoidance experiment in semi-autonomous mode demonstrated the algorithms’reliability and quickness of the obstacle detection and obstacle avoidance. The floor unevenness experiment and collision avoidance experiment proved the safe-driving of smart wheelchair. The wall following experiment in autonomous mode showed the effectiveness of the smart wheelchair’s multi-behaviors fusion algorithm. |