A representation method based on the probability of collision for safe robot navigation in domestic environments

Abstract

This paper introduces a three-dimensional volumetric representation for safe navigation. It is based on the OctoMap representation framework that probabilistically fuses sensor measurements to represent the occupancy probability of volumes. To achieve safe navigation in a domestic environment this representation is extended with a model of the occupancy probability if no sensor measurements are received, and a proactive approach to deal with unpredictably moving obstacles that can arise from behind occlusions by always expecting obstacles to appear on the robot's path. By combining the occupancy probability of volumes with the position uncertainty of the robot, a probability of collision is obtained. It is shown that by relating this probability to a safe velocity limit a robot in a real domestic environment can move close to a certain maximum velocity but decides to attain a slower safe velocity limit when it must, analogous to velocity limits and warning signs in traffic.

Collisions as Information Sources in Densely Packed Multi-Robot Systems Under Mean-Field Approximations

“…The path planning literature contains several ways in which the probability of collisions can be calculated under motion and sensory uncertainty e.g., [6,8,22]. In many cases, Monte Carlo based techniques have been used to obtain estimates of the collision probabilities, e.g., [16], while the majority of the analytical techniques developed compute the probability of collision along a specific path taken through the workspace in the presence of stationary or moving obstacles, e.g., [15,30].…”

Computationally Efficient Obstacle Avoidance Trajectory Planner for UAVs Based on Heuristic Angular Search Method

“…In the previous work [5] , they use Octomap build on point cloud raw data to develop their own method and gained good experimental results. In another way, in order to reduce the time consumed by this step of building the map, some researchers have directly planned the trajectory on the original point cloud.…”

“…The optimization problem is defined in (5), where the subscript n presents the current step in a rolling process of the whole planner, p start is the position of the drone when the planner starts to work [18]. v max and a max are the kinematic constraints for speed and acceleration respectively, t max is the upper bound for the time which can be used to finish the predicted piece of trajectory.…”

A representation method based on the probability of collision for safe robot navigation in domestic environments

“…First, it will be described how this representation, based on the OctoMap framework (Hornung et al 2013) and first introduced in Coenen et al (2014), uses probabilistic fusion of sensor measurements to be robust against uncertainty in sensing. Secondly, it is described how uncertainty due to unknown space is represented if no measurements are received and how the probability of obstacles appearing on the robot's path from behind occlusions is represented using a proactive approach (Alami et al 2002).…”