Understanding human interaction for probabilistic autonomous navigation using Risk-RRT approach

Abstract

Abstract-With the growing demand of personal assistance to mobility and mobile service robotics, robot navigation systems must be "aware" of the social conventions followed by people. They must respect proximity constraints but also respect people interacting. For example, they may not break interaction between people talking, unless the occupants want to take part in the conversation. In this case, they must be able to join the group using a socially adapted behavior. This paper proposes a risk-based navigation method including both the traditional notion of risk of collision and the notion of risk of disturbance. Results exhibit new emerging behavior showing how a robot takes into account social conventions in its navigation strategy.

Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation

“…Further, the algorithm is implemented in simulation only, and the density of humans in the simulated robotic workspace is kept quite low (approximately 0.1 person/m 2 ). Rios-Martinez et al (2011) adopted a ''human-centric'' approach as well, but instead of using the proxemic rules of Hall (1966), they use the criteria of Lam et al (2011) instead. They incorporate these rules of personal space into the robot's behavior by extending the Risk-RRT algorithm developed by Fulgenzi et al (2009).…”

Adaptive spacing in human-robot interactions

“…The goal of these experiments is to demonstrate the utility of the proposed framework in diverse scenarios. Following up earlier work [31], [20] that (a) (b) developed socially-intelligent path planning strategies on the basis of a given set of interacting people our focus here is on the actual extraction of the higher-level spatial relationships within real human-human and human-robot encounters 1 . a) Experiment #1: Two trials were first performed by each of three participants in a scenario where the human and the robot approach each other along a narrow corridor.…”

“…Although these depend on a variety of factors such as cultural background, age, etc. they are statistically biased towards the right [9], [16], [17] and frontal area [18], [19], [20] respectively. Otherwise, in the absence of such bias the shape of social space is reduced to a kernel in the form of concentric spheres or ellipses.…”

Experiments in Leader Classification and Following with an Autonomous Wheelchair

“…Several successful techniques of motion planning address the problem of navigation in populated environments using probabilistic and predictive approaches [1], [2] and on models of social interactions [3]. The idea is that, by estimating areas that ought to be occupied in the future, these algorithms can create collision-free motion plans that may also respect social conventions.…”