Sensorless collision detection and control by physical interaction for wheeled mobile robots

Abstract

In this paper, we present the adaptation of a sensorless (in De Luca's sense [1], i.e., without the use of extra sensors,) collision detection approach previously used on robotic arms to mobile wheeled robots. The method is based on detecting the torque disturbance and does not require a model of the robot's dynamics. We then consider the feasibility of developing control by physical interaction strategies using the described adapted technique.

Comparison of Interaction Modalities for Mobile Indoor Robot Guidance: Direct Physical Interaction, Person Following, and Pointing Control

“…However, these methods limit the surface on the robot body where physical interaction can take place. Doisy [14] has presented a proof of concept that an indoor mobile robot can be controlled by applying force to any part of the robot body without the need of external sensors. Based on this concept, in this paper we developed a new DPI interaction modality that does not rely on external sensors.…”

“…The study was performed for two guidance tasks that differed in the required accuracy of movement. The three interaction modalities developed and evaluated in this paper were based on [14], [15], [16] implementing necessary insights and modifications to further advance these interaction modalities. The interaction modalities described in this section were compared in a study in an apartment facility with a total of 24 users from different backgrounds.…”

“…Natural interfaces aim to create a seamless interaction between the human and the machine [13]. Several advanced natural interfaces for control of robot"s motions along a designated path using physical interaction [14], person-following [15] and pointing [16], which represent different interaction modalities have never been evaluated and compared in user studies in the context of robot guidance. It is important to evaluate each interaction modality within the context of its uses and provide quantitative results in the comparison [2].…”

“…They were chosen and designed with the goal to make personal robots simple to use in a home environment for untrained users. Extensions of [14], [15], [16] include new developments. While DPI modality investigated a tactile human-robot interaction without the use of additional sensors, person-following and pointing-control modalities exploited a powerful human tracking ability of a recently released depth camera.…”

Full-body collision detection and reaction with omnidirectional mobile platforms: a step towards safe human–robot interaction

“…One simple way to detect external forces is by comparing actual and desired positions [12] or velocities [4]. This method is easy to implement because it can use the built-in encoders on the robot joints or wheels to detect external forces.…”