Safe navigation and active vision for autonomous vehicles: A purposive and direct solution

Two capabilities strongly desired in the operation of autonomous vehicles are safe navigation and active vision. It is a challenge to develop autonomous vehicles capable of operating in complicated, unpredictable, and hazardous environments. To navigate autonomous vehicles safely, obstacles such as protrusions, depressions, and steep terrains must be discriminated from terrain before any path planning and obstacle avoidance activity is undertaken. In the first part of the dissertation, a purposive and direct approach to obstacle detection for safe navigation has been developed. The method finds obstacles in a 2-D image-based space, as opposed to 3-D reconstructed space, using optical flow. The theory derives from new visual linear invariants based on optical flow. Employing the linear invariance property, obstacles can be directly detected by using a reference flow line obtained from measured optical flow. The main features of this approach are that (1) 2-D visual information (i.e., optical flow) is directly used to detect obstacles; no range, 3-D motion, or 3-D scene geometry is recovered; (2) the method finding protrusions and depressions is valid for the vehicle (or camera) undergoing general six-degree-of-freedom motion; (3) the error sources involved are reduced to a minimum, since the only information required is one component of optical flow. A number of experiments from both synthetic and real image data suggest that the approach is effective and robust. The method is demonstrated on both ground and air vehicles.; To enhance the capability of active vision, the computer controlled camera system must be calibrated. In the second part of this dissertation, a purposive and direct method for the calibration of the camera system is developed. The method calibrates all the joint poses and link frames of the manipulator and the camera poses related to the same world frame. The solution for the manipulator, camera-to-manipulator and base-to-world calibration is developed in closed form. The main features of this technique are that (1) only the form of one-equation-solving-one-parameter is used; (2) the camera-to-manipulator and base-to-world calibration are the by-products of the manipulator calibration procedure.