Path Prediction Of Guide For Intelligent Wheelchair

As the progress of technology and the development of our society, it isforeseeable that robots will play an important role in our future daily life. In order tofinish daily jobs, robots need the ability to accompany human beings by side. Thispaper proposes a way of human route prediction for Intelligent Wheelchair based onNeural Network algorithm and smoothing spline function, providing a base for therobot to move in parallel with the target human while keeping constant distance.Neural Network is a developed algorithm that has been used successfully in manyfields. Because of its ability to learn unknown, non-linear functions, Neural Networkhas become a powerful tool on research fields like pattern recognition, identificationand prediction. Spline function is naturally smooth in its domain, which means that ithas the first and second derivative at every point in its domain, and they are continues.Such properties of spline function make it especially useful to provide robot a smoothpath to follow.This paper introduces these two methods, and designs a path extraction methodfrom raw data of human footprints. The learning method of Neural Network andSmoothing Spline is adopted to predict future human route.