Improved Autonomous Navigation Through Optimal Sensor Outliers

One of the greatest challenges associated with pose estimation using cameras or sensors such as LIDAR is removal of sensor outliers.nbsp; Outliers exist for numerous reasons, such as incorrectly associated feature points, extraneous data points, etc.nbsp; Without proper removal of outliners, an incorrect pose estimate may result, which can lead to difficulty in achieving mission goals for autonomous operations.nbsp; A common approach to remove outlines is the Random Sample Consensus (RANSAC) algorithm, which is an iterative and non-deterministic algorithm.nbsp; RANSAC may not be optimal, especially if an incorrect model is used, thus leading to outliers that may pass the RANSAC test but may degrade the pose estimate.nbsp;The XAnalytix Systems team has developed an optimal closed-form approach to replace the RANSAC algorithm.nbsp; It is optimal in that it is based on using the statistical properties of the sensor error in its derivation.nbsp; The heart of the solution is based on an optimally derived pose estimation solution from a Total Least Squares (TLS) approach.nbsp; A byproduct of the TLS solution is the error-covariance of the sensor residuals.nbsp; The error-covariance is the key to remove outliers.nbsp; Because the error-covariance is optimal, it is believed that using it will result in a more robust approach to remove outliers than the standard RANSAC algorithm and its variants. The proposed effort will focus on studying the effectiveness of the newly derived closed-form error-covariance in a new RANSAC-type algorithm, called the Statistical Optimal RANSAC (SO-RANSAC) algorithm.nbsp;It is expected that at the completion of Phase I the optimal nature of the SO-RANSAC algorithm, compared to traditional RANDAC-type algorithms, will be verified through simulation testing within a realistic test environment. nbsp;This initial testing and hardware configuration will be used to expedite prototyping and system tests to be conducted during Phase II.