|
Abstract: Signal processing system of the new robotics sensitive skin
based-on DSP and CPLD is described in detail. Parallel signal processing
is implemented by using modularized structure and high-speed float DSP
and CPLD. For the multi-sensor data fusion of sensitive skin, an on-line
adaptive weighted fusion algorithm is realized. The algorithm can adjust
the fused sensor’s weight according to change in the sensor’s variance.
At last, keep the square error of fusion system at the minimum. The
algorithm can adapt itself to the changes of sensor’s character. The
experimental result shows that signal processing system of the sensitive
skin can real-timely provide environment information for obstacle
avoidance of multi-joint robot within sensory area.
Key words: Sensitive skin DSP and CPLD Parallel signal processing Data
fusion Adaptive weighted fusion algorithm
CLC No: TP24
国家自然科学基金(50105002)和哈尔滨工业大学基金(HIT.2001.14)资助项目.
Received 20040812, received in revised form 20041206
|