A Real-Time Approach for Chance-Constrained Motion Planning With Dynamic Obstacles

Abstract

Uncertain dynamic obstacles, such as pedestrians or vehicles, pose a major challenge for optimal robot navigation with safety guarantees. Previous work on motion planning has followed two main strategies to provide a safe bound on an obstacle's space: a polyhedron, such as a cuboid, or a nonlinear differentiable surface, such as an ellipsoid. The former approach relies on disjunctive programming, which has a relatively high computational cost that grows exponentially with the number of obstacles. The latter approach needs to be linearized locally to find a tractable evaluation of the chance constraints, which dramatically reduces the remaining free space and leads to over-conservative trajectories or even unfeasibility. In this work, we present a hybrid approach that eludes the pitfalls of both strategies while maintaining the original safety guarantees. The key idea consists in obtaining a safe differentiable approximation for the disjunctive chance constraints bounding the obstacles. The resulting nonlinear optimization problem is free of chance constraint linearization and disjunctive programming, and therefore, it can be efficiently solved to meet fast real-time requirements with multiple obstacles. We validate our approach through mathematical proof, simulation and real experiments with an aerial robot using nonlinear model predictive control to avoid pedestrians.

A Survey of Computer Vision Methods for 2D Object Detection from Unmanned Aerial Vehicles

“…Examples of tasks and applications of UAVs that require situational awareness, where the use of an object detector is either essential or it substantially improves its performance (see Figure 2 ), are the following: Path planning or obstacle avoidance: consist on the generation of collision-free trajectories and motion commands based on the available knowledge of the existing obstacles of the environment [ 36 ]. Although it can be based on pure metric maps [ 43 ], its performance can be increased when having semantic information such as the kind of objects of the environment [ 44 , 45 ]. Object or person tracking (following or servoing): consist on the generation of trajectories and motion commands that ensure that a particular targeted object (or person) are followed while they move [ 46 , 47 ].…”

A Dynamic Path Planning Method for Social Robots in the Home Environment

“…As the pivotal foundation in social robot navigation, a wide variety of path planning methods have been proposed for robots navigating in different environments [7][8][9][10]. In general, these methods can be divided into global and local methods according to the completeness of map information known before path planning progress.…”

“…The map size is 640 × 480. The starting point of the robot is (10,10). The whole process of the robot moving from the initial point to the destination is simulated as one loop.…”

“…The size of map 2 is 640 × 480. The starting point of the robot is (10,10). The moving trajectories of obstacles and destinations are designed according to the new map.…”

Evaluation of Estimators for Hybrid GNSS-Terrestrial Localization in Collaborative Networks

“…The current development and research efforts in 5G cellular networks suggest that future cellular user equipment will feature device-to-device ranging capabilities on top an improved device-to-infrastructure ranging features [14]- [18]. The combined use of device-to-device and deviceto-infrastructure ranging can provide a positioning solution whose accuracy and reliability depends on the density of devices and infrastructure elements.…”