Method and system of integrity monitoring for visual odometry

A system comprising:

• at least one set of stereo vision sensors onboard a vehicle;
• and
• a processer onboard the vehicle, the processor operatively coupled to the stereo vision sensors, the processor operative to perform a method of integrity monitoring for visual odometry, the method comprising;
• extracting a plurality of features from each of a first image and a second image captured by the stereo vision sensors, wherein the first image is captured at a first time epoch and the second image is captured at a second time epoch;
• performing a temporal feature matching process operative to match the extracted features from the first image at the first time epoch with the extracted features from the second image at the second time epoch, wherein the temporal feature matching process uses a feature mismatching limiting discriminator to reduce a likelihood of feature mismatch, between features in the first and second images, to below a first predetermined threshold;
• performing a range, or depth, recovery process operative to provide stereo feature matching between two images respectively taken by the stereo vision sensors at the same time epoch, wherein the range, or depth, recovery process uses a range error limiting discriminator to limit features having range errors greater than a second predetermined threshold;
• performing an outlier rejection process using a modified random sample consensus (RANSAC) technique operative to limit feature moving events that are greater than a third predetermined threshold;
• characterizing a feature error magnitude and feature fault probabilities using overbounding Gaussian models;
• and
• performing a state vector estimation process with integrity check using a solution separation technique to determine a change in rotation and a change in translation between the first image and the second image, to determine error statistics of the change in rotation and the change in translation between the first image and the second image, to detect any faults, and to compute a protection level or integrity risk.