All information about the crowdbot challenge is on this webpage.
This repository contains the implementation of DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data.
Map-matcher is a ros node which matches a source map to a reference map.
This repository contains tools for active SLAM in crowded environments. It contains code that has been tested and used with a pioneer in simulation, a real pepper robot and a real turtlebot3-pi.
Redirecting Driver Support (RDS) is a method for robots to reactively avoid imminent collisions with moving objects.
High level multi-behaviour planning for navigating through different crowd scenarios.
Simulator and benchmark implementations of RL-based robot navigation algorithms.
Compliance control for mobile robots to deal safely with impact through a sliding response, advancing around pedestrians/obstacles in closed-loop force control.
This dataset contains injury measures during collisions between a mobile service robot – Qolo – and pedestrian dummies: male adult Hybrid-III (H3) and child model 3-years-old (Q3). We present multiple collision scenarios for the assessment of pedestrian safety, considering possible impacts at the legs for adult pedestrians, and legs, chest and head for children. The dataset is available here DOI:10.5281/zenodo.5266447, and the data processing repository on GitHub.
This dataset contains over 250k frames of robot navigation in raw crowds in the city of Lausanne, Switzerland with the personal mobility robot Qolo in semi-autonomous navigation. It includes egocentric sets of frontal and rear 3D point clouds from Velodyne VLP-16 and labelled RGBD videos. The dataset is available here DOI:10.21227/ak77-d722, and data processing repository on GitHub.Funded by the EU CrowdBot project.