Avoiding moving obstacles

Abstract

To successfully move our hand to a target, we must consider how to get there without hitting surrounding objects. In a dynamic environment this involves being able to respond quickly when our relationship with surrounding objects changes. People adjust their hand movements with a latency of about 120 ms when the visually perceived position of their hand or of the target suddenly changes. It is not known whether people can react as quickly when the position of an obstacle changes. Here we show that quick responses of the hand to changes in obstacle position are possible, but that these responses are direct reactions to the motion in the surrounding. True adjustments to the changed position of the obstacle appeared at much longer latencies (about 200 ms). This is even so when the possible change is predictable. Apparently, our brain uses certain information exceptionally quickly for guiding our movements, at the expense of not always responding adequately. For reaching a target that changes position, one must at some time move in the same direction as the target did. For avoiding obstacles that change position, moving in the same direction as the obstacle is not always an adequate response, not only because it may be easier to avoid the obstacle by moving the other way, but also because one wants to hit the target after passing the obstacle. Perhaps subjects nevertheless quickly respond in the direction of motion because this helps avoid collisions when pressed for time.

Ultra-fast selection of grasping points

“…Why did our subjects respond to the rotation of the ball at all? An interpretation that would be consistent with earlier studies (Day and Lyon 2000; Pisella et al 2000;Aivar et al 2008;van Mierlo et al 2009) is that the initial response was an automatic reaction to motion of the selected points on the surface of the ball. The response was aborted after an additional 40 ms.…”

“…Consequently, when confronted with a moving obstacle during a goal-directed movement, people initially follow the motion of the obstacle. If doing so is inappropriate, they correct the response ϳ50 ms later (Aivar et al 2008). Similarly, when instructed to respond to a target jump by moving in the opposite direction than the target, people briefly follow the target before moving in the opposite direction (Day and Lyon 2000).…”

Learning robotic eye–arm–hand coordination from human demonstration: a coupled dynamical systems approach

“…In the study by Paulignan et al (1991), subjects were able to instantly modulate, by relying on visual feedback, the arm and hand movements with respect to online perturbations of the position and shape of the target object, with only minimal increase in the response time (∼100 ms) compared with motion in the absence of perturbations. Aivar et al (2008) studied adjustments of the hand movements with respect to abrupt online perturbations of obstacles and/or the target. They found similar latencies to those reported by Paulignan et al (1991) for the responses to the perturbations of the target position and slightly longer adaptation latencies for the obstacles.…”

“…Rothkopf and Ballard (2009), who studied human navigation in an immersed graphic environment, reported that subjects fixate the edges of obstacles for the purpose of planning a walking path for obstacle avoidance. Aivar et al (2008) provided evidence that fast arm responses to the displacement of obstacles are triggered by a reaction to retinal motion of moving obstacles. Many important questions still remain unanswered.…”

“…Vision is particularly useful to plan motion so as to avoid obstacles without touching them (Johansson et al 2001). It also enables to react rapidly in the face of a sudden perturbation, such as an obstacle entering the workspace (Aivar et al 2008). There is a tight coupling between visual and motor modalities when driving prehensile motion (Prablanc et al 1979;Land et al 1999;Johansson et al 2001).…”