Mission-Adaptive Crowd Navigation For Mobile Robots
Human-robot/agent interaction; Simulation; Social models
Recent developments in mobile robotics made feasible the near future scenario of mobile robots assisting individual persons. Such robots must maintain a sufficient distance from their human owners to be able to offer assistance, but otherwise they need to be inconspicuous and observe the prevailing social and cultural norms. We are considering a scenario of mobile robots assisting a peacekeeper soldier patrolling a market with a dense crowd. The robot must balance the costs related to its mission (the danger of loosing contact with its owner) and the social cost of violating the crowd members' personal space. We develop a technique through which we predict the mission cost of different decisions, and use it to adapt the robot's strategies for resolving the micro-conflicts encountered in crowd navigation. We show that this adaptive strategy outperforms strategies of consistent politeness / assertiveness over a variety of scenarios.
13th International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2014
Number of Pages
Article; Proceedings Paper
Source API URL
Arif, Saad; Khan, Saad Ahmad; and Bölöni, Ladislau, "Mission-Adaptive Crowd Navigation For Mobile Robots" (2014). Scopus Export 2010-2014. 9173.