We present methods for efficiently maintaining human head orientation using low-cost MEMS sensors. We particularly address gyroscope integration and compensation of dead reckoning errors using gravity and magnetic fields. Although these problems have been well-studied, our performance criteria are particularly tuned to optimize user experience while tracking head movement in the Oculus Rift Development Kit, which is the most widely used virtual reality headset to date. We also present novel predictive tracking methods that dramatically reduce effective latency (time lag), which further improves the user experience. Experimental results are shown, along with ongoing research on positional tracking.

This paper defines and analyzes a simple robot with local sensors that moves in an unknown polygonal environment. The robot can execute wall-following motions and can traverse the interior of the environment only when following parallel to an edge. The robot has no global sensors that would allow precise mapping or localization. Special information spaces are introduced for this particular model. Using these, strategies are presented to solve several tasks: 1) counting vertices, 2) computing the path winding number, 3) learning a combinatorial map, which is called the cut ordering, that encodes partial geometric information, and 4) solving pursuit-evasion problems.

The cost of nearest-neighbor (NN) calls is one of the bottlenecks in the performance of sampling-based motion-planning algorithms. Therefore, it is crucial to develop efficient techniques for NN searching in configuration spaces arising in motion planning. In this paper, we present and implement an algorithm for performing NN queries in Cartesian products of R, S1, and RP3, the most common topological spaces in the context of motion planning. Our approach extends the algorithm based on kd-trees, called ANN, developed by Arya and Mount for Euclidean spaces. We prove the correctness of the algorithm and illustrate substantial performance improvement over the brute-force approach and several existing NN packages developed for general metric spaces. Our experimental results demonstrate a clear advantage of using the proposed method for both probabilistic roadmaps and rapidly exploring random trees

Sampling-based planners have solved difficult problems in many applications of motion planning in recent years. In particular, techniques based on the Rapidly-exploring Random Trees (RRTs) have generated highly successful single-query planners. Even though RRTs work well on many problems, they have weaknesses which cause them to explore slowly when the sampling domain is not well adapted to the problem. In this paper we characterize these issues and propose a general framework for minimizing their effect. We develop and implement a simple new planner which shows significant improvement over existing RRT-based planners. In the worst cases, the performance appears to be only slightly worse in comparison to the original RRT, and for many problems it performs orders of magnitude better.

Sampling based planners have become increasingly efficient in solving the problems of classical motion planning and its applications. In particular, techniques based on the rapidly-exploring random trees (RRTs) have generated highly successful single-query planners. Recently, a variant of this planner called dynamic-domain RRT was introduced by Yershova et al. (2005). It relies on a new sampling scheme that improves the performance of the RRT approach on many motion planning problems. One of the drawbacks of this method is that it introduces a new parameter that requires careful tuning. In this paper we analyze the influence of this parameter and propose a new variant of the dynamic-domain RRT, which iteratively adapts the sampling domain for the Voronoi region of each node during the search process. This allows automatic tuning of the parameter and significantly increases the robustness of the algorithm. The resulting variant of the algorithm has been tested on several path planning problems.

This paper addresses the problem of generating uniform deterministic samples over the spheres and the three-dimensional rotation group, SO(3). The target applications include motion planning, optimization, and verification problems in robotics and in related areas, such as graphics, control theory and computational biology. We introduce an infinite sequence of samples that is shown to achieve: 1) low-dispersion, which aids in the development of resolution complete algorithms, 2) lattice structure, which allows easy neighbor identification that is comparable to what is obtained for a grid in /spl Ropf//sup d/, and 3) incremental quality, which is similar to that obtained by random sampling. The sequence is demonstrated in a sampling-based motion planning algorithm.