Huili Yu

Cooperative Path Planning for Target Tracking in Urban Environments Using Unmanned Air and Ground Vehicles

As the need for autonomous reconnaissance and surveillance missions in cluttered urban environments has been increasing, this paper describes a cooperative path planning algorithm for tracking a moving target in urban environments using both unmanned air vehicles (UAVs) and unmanned ground vehicles (UGVs). The novelty of the algorithm is that it takes into account vision occlusions due to obstacles in the environment. The algorithm uses a dynamic occupancy grid to model the target state, which is updated by sensor measurements using a Bayesian filter. Based on the current and predicted target behavior, the path planning algorithm for a single vehicle (UAV/UGV) is first designed to maximize the sum of the probability of detection over a finite look-ahead horizon. The algorithm is then extended to multiple vehicle collaboration scenarios, where a decentralized planning algorithm relying on an auction scheme is designed to plan finite look-ahead paths that maximize the sum of the joint probability of detection over all vehicles.

Probabilistic path planning for cooperative target tracking using aerial and ground vehicles

In this paper, we present a probabilistic path planning algorithm for tracking a moving ground target in urban environments using UAVs in cooperation with UGVs. The algorithm takes into account vision occlusions due to obstacles in the environments. The target state is modeled using the dynamic occupancy grid and the probability of the target location is updated using Bayesian Altering. Based on the probability of the targets current and predicted locations, the path planning algorithm is designed to generate paths for a single UAV or UGV maximizing the sum of probability of detection over a finite look-ahead. For target tracking using multiple vehicle collaboration, a decentralized planning algorithm using an auction scheme generates paths maximizing the sum of joint probability of detection over the finite look ahead horizon. Simulation results show the proposed algorithm is successful in solving the target tracking problem in urban environments.

Observability-based local path planning and collision avoidance for micro air vehicles using bearing-only measurements

In this paper we detail an observability based path planning algorithm for Small and Miniature Air Vehicles (MAVs) navigating among multiple static obstacles. Bearing only measurements are utilized to estimate the time-to-collision (TTC) and bearing to obstacles using the extended Kalman filter (EKF). For the error covariance matrix computed by the EKF to be bounded, the system should be observable. We perform a nonlinear observability analysis to obtain the necessary conditions for complete observability. We use these conditions to design a path planning algorithm which simultaneously minimizes the uncertainties in state estimation while avoiding collisions with obstacles. Simulation results show that the planning algorithm successfully solves the single and multiple obstacle avoidance problems for MAVs while improving the estimation accuracy.

Vision-based local multi-resolution mapping and path planning for Miniature Air Vehicles

Miniature Air Vehicles (MAVs) are often used for low altitude flights where unknown obstacles might be encountered. Path planning and obstacle avoidance for MAVs involve planning a feasible path from an initial state to a goal state while avoiding obstacles in the environment. This paper presents a vision-based local multi-resolution mapping and path planning technique for MAVs using a forward-looking onboard camera. A depth map, which represents the time-to-collision (TTC) and bearing information of the obstacles, is obtained by computer vision algorithms. To account for measurement uncertainties introduced by the camera, a multi-resolution map in the body frame of the MAV is created in polar coordinates. Using the depth map, the locations of the obstacles are determined in the multi-resolution map. Dijkstras algorithm is employed to find a collision-free path in the body frame. The simulation and flight test results show that the proposed technique is successful in solving path planning and multiple obstacles avoidance problems for MAVs.

Observability-based local path planning and obstacle avoidance using bearing-only measurements

In this paper we present an observability-based local path planning and obstacle avoidance technique that utilizes an extended Kalman Filter (EKF) to estimate the time-to-collision (TTC) and bearing to obstacles using bearing-only measurements. To ensure that the error covariance matrix computed by an EKF is bounded, the system should be observable. We perform a nonlinear observability analysis to obtain the necessary conditions for complete observability of the system. These conditions are used to explicitly design a path planning algorithm that enhances observability while simultaneously avoiding collisions with obstacles. We analyze the behavior of the path planning algorithm and specially define the environments where the path planning algorithm will guarantee collision-free paths that lead to a goal configuration. Numerical results show the effectiveness of the planning algorithm in solving single and multiple obstacle avoidance problems while improving the estimation accuracy.

Vision-based Three Dimensional Navigation Frame Mapping and Planning for Collision Avoidance for Micro Air Vehicles

Path planning and obstacle avoidance involve planning a feasible path from the initial to goal conflgurations while avoiding obstacles. This paper presents a vision-based three dimensional navigation frame mapping and path planning technique for collision avoidance for Micro Air Vehicles (MAVs) using a forward-looking onboard camera. Using computer vision algorithms, a depth map representing the range and bearing to obstacles is obtained. Based on the depth map, we use the extended Kalman Filter (EKF) to estimate the range and azimuth to obstacles and the height of obstacles and use the joint compatibility branch and bound (JCBB) algorithm to associate the camera measurements with the existing obstacles. A three dimensional map, constructed in cylindrical coordinates, is then created in the navigation frame of the MAV. We employ the Rapidly-Exploring Random Tree (RRT) algorithm to flnd collision-free Dubins paths in the navigation frame. The simulation results show that the proposed technique is successful in solving path planning and multiple obstacles avoidance problems for MAVs.

Uncertainty Estimation for KLT Tracking

The Kanade-Lucas-Tomasi tracker (KLT) is commonly used for tracking feature points due to its excellent speed and reasonable accuracy. It is a standard algorithm in applications such as video stabilization, image mosaicing, egomotion estimation, structure from motion and Simultaneous Localization and Mapping (SLAM). However, our understanding of errors in the output of KLT tracking is incomplete. In this paper, we perform a theoretical error analysis of KLT tracking. We first focus our analysis on the standard KLT tracker and then extend it to the pyramidal KLT tracker and multiple frame tracking. We show that a simple local covariance estimate is insufficient for error analysis and a Gaussian Mixture Model is required to model the multiple local minima in KLT tracking. We perform Monte Carlo simulations to verify the accuracy of the uncertainty estimates.

Improving cooperative tracking of an urban target with target motion model learning

Tracking a ground urban target with multiple unmanned aerial vehicles (UAVs) is a challenging problem due to cluttered urban environments and coordination of nonholonomic UAV motion. Our previous work has demonstrated in simulation that machine learning can be used in such an environment to learn a model of target motion and thereby improve tracking performance. We extend this previous work by creating a more realistic simulation using road network and building height data extracted from downtown San Diego. We demonstrate effectiveness of target motion model learning in the new simulation environment. Additionally, we demonstrate performance improvement by extending the algorithm used to coordinate the UAVs for tracking the urban target.

Uncertainty estimation for random sample consensus

The RANdom SAmple Consensus (RANSAC) algorithm, as a robust parameter estimator, has been widely used to remove gross errors. However, there is less work on analyzing the uncertainty produced by the RANSAC. This paper fills this gap by presenting an uncertainty estimation algorithm for the RANSAC. Based on a thorough analysis on the uncertainty of the model parameters generated during the random hypothesis sampling process of the RANSAC, we derive the probability that each hypothesis is selected as the best hypothesis by the RANSAC. Using the probability of the best hypothesis, we characterize the error expectation and error covariance of the model parameter estimates and compute the probability of each data point being an inlier. Three models including line fitting, homography, and essential matrix are used to evaluate the performance of the uncertainty estimation algorithm. Results demonstrate that the uncertainty produced by the RANSAC is characterized successfully by the proposed algorithm.

Vision-based local-level frame mapping and planning in spherical coordinates for Miniature Air Vehicles

This paper presents a vision-based collision avoidance technique for small and miniature air vehicles (MAVs) using local-level frame mapping and planning in spherical coordinates. To explicitly address the obstacle initialization problem, the maps are parameterized using the inverse time-to-collision (TTC). Using bearing-only measurements, an extended Kalman filter is employed to estimate the inverse TTC, azimuth, and elevation to obstacles. Nonlinear observability analysis is used to derive conditions for the observability of the system. Based on these conditions, we design a path planning algorithm that simultaneously minimizes the uncertainties in state estimation while avoiding collisions with obstacles. The behavior of the planning algorithm is analyzed, and the characteristics of the environment in which the planning algorithm is guaranteed to generate collision-free paths for MAVs are described. Numerical results show that the proposed method is successful in solving the path planning problem for MAVs.