Tag Archives: Parallelization Of Robotic Algorithms

Hardware efficient collision avoidance for mobile robots through the use of interval arithmetics and parallelism

Pranjal Vyas, Leena Vachhani, K Sridharan, Hardware-efficient interval analysis based collision detection and avoidance for mobile robots. Mechatronics, Volume 62, 2019, DOI: 10.1016/j.mechatronics.2019.102258.

Collision detection and avoidance is challenging when the mobile robot is moving among multiple dynamic obstacles. A hardware-efficient architecture supporting parallel implementation is presented in this work for low-power, faster and reliable collision-free motion planning. An approach based on interval analysis is developed for designing an efficient hardware architecture. The proposed architecture achieves parallelism which can be combined with any robotic task involving multiple obstacles. Interval arithmetic is used for representing the pose of the robot and the obstacle as velocity intervals in a fixed time period. These intervals correspond to sub-intervals such as arcs and line-segments. In particular, the collision detection problem for dynamic objects involves the computation of line segment-arc intersections and segment-segment intersections. The intersection of these boundary curves is carried out in a hardware-efficient manner so that it avoids complex arithmetic computations such as multiplication, division etc and exploits parallelism. We develop several results on intersection of these sub-intervals for collision detection and use them to obtain a hardware-efficient collision detection algorithm that requires only shift and add-type of computations. The algorithm is further used in developing a hardware-efficient technique for finding an exhaustive set of solutions for avoiding collision of the robot with dynamic obstacles. Simulation results in MatLab and experiments with a Field Programmable Gate Array (FPGA)-based robot show that a variety of collision avoidance techniques can be implemented using the proposed solution set that guarantees collision avoidance with multiple obstacles.

Massive parallelization of POMDPs with a very good state-of-the-art review

Taekhee Lee, Young J. Kim (2015), Massively parallel motion planning algorithms under uncertainty using POMDP , The International Journal of Robotics Research, Vol 35, Issue 8, pp. 928 – 942, DOI: 10.1177/0278364915594856.

We present new parallel algorithms that solve continuous-state partially observable Markov decision process (POMDP) problems using the GPU (gPOMDP) and a hybrid of the GPU and CPU (hPOMDP). We choose the Monte Carlo value iteration (MCVI) method as our base algorithm and parallelize this algorithm using the multi-level parallel formulation of MCVI. For each parallel level, we propose efficient algorithms to utilize the massive data parallelism available on modern GPUs. Our GPU-based method uses the two workload distribution techniques, compute/data interleaving and workload balancing, in order to obtain the maximum parallel performance at the highest level. Here we also present a CPU–GPU hybrid method that takes advantage of both CPU and GPU parallelism in order to solve highly complex POMDP planning problems. The CPU is responsible for data preparation, while the GPU performs Monte Cacrlo simulations; these operations are performed concurrently using the compute/data overlap technique between the CPU and GPU. To the best of the authors’ knowledge, our algorithms are the first parallel algorithms that efficiently execute POMDP in a massively parallel fashion utilizing the GPU or a hybrid of the GPU and CPU. Our algorithms outperform the existing CPU-based algorithm by a factor of 75–99 based on the chosen benchmark.