Brian R. Geiger

Optimal Path Planning of UAVs Using Direct Collocation with Nonlinear Programming

A trajectory generation algorithm using direct collocation with nonlinear programming is successfully demonstrated in simulation. Direct collocation, which approximates the states and controls with piecewise polynomials, has been widely used in space and manned aircraft applications, but has only seen limited use in UAV applications. The algorithm is successfully applied to the generation of a UAV trajectory that provides maximum viewing time for a camera mounted on the UAV. The target can be stationary or moving. Multiple UAVs are considered. In this case, the objective is to provide maximum sensor coverage time using a combination of the UAVs. No specific initial guesses are required to ensure the algorithm is successful.

Flight Testing a Real-Time Direct Collocation Path Planner

A path-planning algorithm using direct collocation with nonlinear programming is demonstrated in both simulation and flight tests. Direct collocation, which approximates the states and controls with piecewise polynomials, has been widely applied in space vehicles and manned aircraft, but has only seen limited use in unmanned aerial vehicle applications. The algorithm is successfully used to generate a path that produces maximal surveillance time of a moving or stationary ground target by a sensor mounted on an unmanned aerial vehicle while compensating for aircraft performance or mission constraints. Flight tests of the path-planning algorithm operating in real time onboard an unmanned aerial vehicle are also presented. These tests include surveilling a stationary and moving target with a video camera while compensating for any wind effects. Additionally, the effect of the use of road data in planning the path is simulated by tracking a second unmanned aerial vehicle flying a predefined pattern.

Intelligent Unmanned Air Vehicle Flight Systems

This paper describes an intelligent autonomous airborne flight capability that is being used as a test bed for future technology development. The unmanned air vehicles (UAVs) fly under autonomous control of both an onboard computer and an autopilot. The onboard computer provides the mission control and runs the autonomous Intelligent Controller (IC) software while the autopilot controls the vehicle navigation and flight control. The autonomous airborne flight system is described in detail. An IC architecture directly applicable to the design of unmanned vehicles is also presented. The UAVs may operate independently or in cooperation with one another to carry out a specified mission. The intelligent UAV flight system is used to evaluate and study autonomous UAV control as well as multi-vehicle collaborative control.

Neural Network Based Trajectory Optimization for Unmanned Aerial Vehicles

A neural network approximation to direct trajectory optimization methods is presented. The method uses neural networks to approximate the dynamics and objective equations integrated over a given time interval. The trajectory is then built recursively and treated as a nonlinear programming problem. The method is compared to a direct collocation method as well as more recent pseudospectral methods and shows competitive results while being computationally faster. In addition, a neural network provides a continuously dierentiable function approximation which may be advantageous when a discontinuous objective function is used in a nonlinear solver. A surveillance trajectory planning problem for an unmanned aerial vehicle is given as an example application and results are presented for all three methods.

Vision-Based Target Geolocation and Optimal Surveillance on an Unmanned Aerial Vehicle

A real-time computer vision algorithm for the identification and geolocation of ground targets was developed and implemented on the Penn State University / Applied Research Laboratory Unmanned Aerial Vehicle (PSU/ARL UAV) system. The geolocation data is filtered using a linear Kalman filter, which provides a smoothed estimate of target location and target velocity. The vision processing routine and estimator are coupled with an onboard path planning algorithm that optimizes the vehicle trajectory to maximize surveillance coverage of the targets. The vision processing and estimation routines were flight tested onboard a UAV system with a human pilot-in-the-loop. It was found that GPS latency had a significant effect on the geolocation error, and performance was significantly improved when using latency compensation. The combined target geolocation and path planning system was tested on the ground using a hardware-in-the-loop simulation, and resulted successful tracking and observation of a fixed target. Timing results showed that is is feasible to implement total system in real time.

Flight Testing a Real Time Implementation of a UAV Path Planner Using Direct Collocation

Flight tests of a path planning algorithm using direct collocation with nonlinear programming (DCNLP) are presented. The path planner operates in real time onboard an unmanned aerial vehicle for these tests. The method plans a path that maximizes the time a target is in view of a camera onboard the aircraft. Tests include surveilling a stationary and moving target while compensating for any wind effects. Additionally, the effect of the use of road data in planning the path is simulated by tracking a second UAV flying a predefined pattern. Finally, a method of commanding the observation of a target from a specific line of sight is presented.

An Intelligent Controller for Collaborative Unmanned Air Vehicles

This paper describes an implementation of an autonomous intelligent controller (IC) architecture for collaborative control of multiple unmanned aerial vehicles (UAVs). Collaborative capabilities include formation flying, search of an area, and cooperative investigation of a target. The IC provides capabilities for sensor data fusion, internal representation of the real-world, and autonomous decision making based on the ICs world model and mission goals. Results of flight tests demonstrating these capabilities are presented. Future work, such as integration of different sensors and collaboration with heterogeneous vehicles, is discussed.

Use of Neural Network Approximation in Multiple-Unmanned Aerial Vehicle Trajectory Optimization

A direct trajectory optimization method that uses neural network approximations is presented. Neural networks are trained to model objective functions, vehicle dynamics, and non-linear constraints. The neural network method reduces computational requirements by removing the need for collocation and providing fast analytical computation of gradients. The method was shown to signicantly reduce computational costs while resulting in trajectories comparable to direct collocation and pseudospectral methods. The method is applied to a multi-aircraft surveillance mission to demonstrate the scalability of the method when adding additional aircraft or objective functions to the optimization problem. Cases are presented in which neural networks can be reused for dierent purposes without additional training.

Use of Neural Network Approximation for Trajectory Optimization of Unmanned Aerial Vehicles with Gimbaled Cameras

A direct trajectory optimization method that uses neural network approximation methods is presented. Neural networks are trained to approximate objective functions and vehicle dynamics. The neural network method reduces computational requirements by removing the need for collocation and providing fast computation of gradients. The method has been shown to significantly reduce computational costs while resulting in trajectories comparable to direct collocation and pseudospectral methods. Since the neural networks readily provide accurate computation of gradients, it removes the need for formulating analytical gradients, so the method is more easily extended to different types of applications with different objective functions and constraints. In previous work, the method was applied towards trajectory optimization of single or multiple UAVs with fixed, downward facing cameras in order to maximize surveillance of a ground target. In this paper the method is extended to optimize trajectories for a UAV equipped with a gimbaled camera, and more complex constraints on the vehicle trajectory are investigated.

Fusion of Unmanned Aerial Vehicle Range and Vision Sensors Using Fuzzy Logic and Particles

This paper presents a novel method for fusing data from a UAV’s range and vision sensors. The range sensor is used to build an elevation map of the flying area. Fuzzy logic is used to detect red barrels in camera images. The world location of a target on the ground is found by fusing the terrain map with image data using both an extended Kalman filter and a particle filter. The target detection system has been tested using images collected onboard a UAV. The terrain mapping system and the fusion system were both tested in simulation.