Obstacle avoidance using image flow in an RT-Linux environment in a PC-104 platform

摘要

Semi-autonomous to fully autonomous robots rely on some form of data collection to operate in their environment. This has traditionally been accomplished using sonar or infra-red sensors to measure the robot's proximity to nearby objects. The most recent of efforts rely on sophisticated sensors, such as LIDAR and stereo vision, and result in solutions which are both expensive economically and computationally. Furthermore, these approaches often provide an overflow of data requiring a great deal of processing. This paper discusses an alternative method using a single camera sampling images periodically to calculate the flow within an image and provide sufficient information to allow a small autonomous robot to navigate a corridor and react to obstacles. This method was implemented on a small robotic platform within an RT-Linux environment. Image data was collected using a CMUCam.