PANTHER: Perception-Aware Trajectory Planner in Dynamic Environments

Abstract

This paper presents PANTHER, a real-time perception-aware (PA) trajectory planner for multirotor-UAVs (Unmanned Aerial Vehicles) in dynamic environments. PANTHER plans trajectories that avoid dynamic obstacles while also keeping them in the sensor field of view (FOV) and minimizing the blur to aid in object tracking. The rotation and translation of the UAV are jointly optimized, which allows PANTHER to fully exploit the differential flatness of multirotors to maximize the PA objective. Real-time performance is achieved by implicitly imposing the underactuated dynamics of the UAV through the Hopf fibration. PANTHER is able to keep the obstacles inside the FOV 7.9 and 1.5 times more than non-PA approaches and PA approaches that decouple translation and yaw, respectively. The projected velocity (and hence the blur) is reduced by 18% and 34%, respectively. This leads to average success rates three times larger than state-of-the-art approaches in multi-obstacle avoidance scenarios. The MINVO basis is used to impose low-conservative collision avoidance constraints in position and velocity space. Finally, extensive hardware experiments in unknown dynamic environments with all the computation running onboard are presented, with velocities of up to 5.8 m/s, and with relative velocities (with respect to the obstacles) of up to 6.3 m/s. The only sensors used are an IMU, a forward-facing depth camera, and a downward-facing monocular camera.

Dynamic smooth and stable obstacle avoidance for unmanned aerial vehicle based on collision prediction

“…The adaptive trajectory optimization method is used to determine the optimal lateral acceleration ⃗ a u . The specific method is expressed by Equations ( 30)- (33).…”

“…Six sample points were used to fit the trajectories. Five points were collected with a sampling interval $$$ \Delta T $$$ of 20 ms. After 400 ms, the sampling point was used as the calibration point for parameter estimation.PLS method 33 …”

“…This scheme can achieve relatively long‐term trajectory predictions; however, the trajectory library may not contain trajectories for all scenarios. In Reference 33, the authors used polynomial fitting to predict obstacle trajectories based on the real‐time perception of obstacle position information. Polynomial curve fitting can achieve long‐distance trajectory fitting; however, the error significantly increases over time.…”

Prioritising paths: An improved cost function for local path planning for UAV in medical applications

“…Most rely on classic deterministic approaches, often based on an optimisation approach [13]. Several algorithms rely on weighted sums as cost functions [14][15][16][17]; however, to the best of the authors' knowledge, no comprehensive investigation of the influence of the type of cost function on the local path planning result is available. Using artificial neural networks and machine learning also becomes increasingly popular for local path planning tasks [18].…”

RAST: Risk-Aware Spatio-Temporal Safety Corridors for MAV Navigation in Dynamic Uncertain Environments

“…Other works that consider dynamic obstacles typically assume that they can be explicitly segmented from static obstacles [6], their number is known and fixed [7], or they can be modeled with given shapes, e.g. ellipsoids [8], [9] and polytopes [10]. However, these assumptions can hardly be satisfied in many real-world scenarios, where an unknown number of arbitrarilyshaped dynamic or static obstacles can appear, and the MAV's sensing and localization contain non-negligible noise.…”