No abstract

To achieve robot navigation in crowded environments having high densities of moving people, it is insufficient to simply consider humans as moving obstacles and avoid collisions with them. That is, the impact of an approaching robot on human movements must be considered as well. Moreover, various navigation methods have been tested in their own environments in the literature, which made them difficult to compare with one another. Thus, we propose an autonomous robot navigation method in densely crowded environments for data-based predictions of robot-human interactions, together with a reproducible experimental test under controlled conditions. Based on localized positional relationships with humans, this method extracts multiple alternative paths, which can implement either following or avoidance, and selects an optimal path based on time efficiency. Each path is selected using neural networks, and the various paths are evaluated by predicting the position after a given amount of time has elapsed. These positions are then used to calculate the time required to reach a certain target position to ensure that the optimal path can be determined. We trained the predictor using simulated data and conducted experiments using an actual mobile robot in an environment where humans were walking around. Using our proposed method, collisions were avoided more effectively than when conventional navigation methods were used, and navigation was achieved with good time efficiency, resulting in an overall reduction in interference with humans. Thus, the proposed method enables an effective navigation in a densely crowded environment, while collecting human-interaction experience for further improvement of its performance in the future.

This paper addresses optimal motion for general machines. Approximation for optimal motion requires a global path planning algorithm that precisely calculates the whole dynamics of a machine in a brief calculation. We propose a path planning algorithm that consists of path searching and pruning algorithms. The pruning algorithmis based on our analysis of state resemblance in general phase space. To confirm precision, calculation cost, optimality and applicability of the proposed algorithm, we conducted several shortest time path planning experiments for the dynamic models of double inverted pendulums. Precision to reach the goal states of the pendulums was better than other algorithms. Calculation cost was 58 times faster at least. We could tune optimality of proposed algorithm via resolution parameters. A positive correlation between optimality and resolutions was confirmed. Applicability was confirmed in a torque based position and velocity feedback control simulation. As a result of this simulation, the double inverted pendulums tracked planned motion under noise while keeping within torque limitations.

Variable Mesh Optimization with Niching (VMO-N) is a framework for multimodal problems (those with multiple optima at several search subspaces). Its only two instances are restricted though. Being a potent multimodal optimizer, the Hill-Valley Evolutionary Algorithm (HillVallEA) uses large populations that prolong its execution. This study strives to revise VMO-N, to contrast it with related approaches, to instantiate it effectively, to get HillVallEA faster, and to indicate methods (previous or new) for practical use. We hypothesize that extra pre-niching search in HillVallEA may reduce the overall population, and that if such a diminution is substantial, it runs more rapidly but effective. After refining VMO-N, we bring out a new case of it, dubbed Hill-Valley-Clustering-based VMO (HVcMO), which also extends HillVallEA. Results show it as the first competitive variant of VMO-N, also on top of the VMO-based niching strategies. Regarding the number of optima found, HVcMO performs statistically similar to the last HillVallEA version. However, it comes with a pivotal benefit for HillVallEA: a severe reduction of the population, which leads to an estimated drastic speed-up when the volume of the search space is in a certain range.

No abstract