Highly paralellizable route planner based on cellular automata algorithms

Stiles, P.; Glickstein, Ira

doi:10.1147/rd.382.0167

Cited by 9 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several parallel path planning algorithms of different categories based on cell maps are published in literature [5], [6], [7] and, like our approach, there already exist algorithms based on CAs [8], [9], [10]. The disadvantage of these approaches is that they have to store significantly more bits per map point than our approach.…”

Section: Related Workmentioning

confidence: 99%

An Optimized FPGA Implementation for a Parallel Path Planning Algorithm Based on Marching Pixels

Schmidt

Fey

2010

2010 International Conference on Reconfigurable Computing and FPGAs

View full text Add to dashboard Cite

Path Planning is one of the most computationally intensive tasks in robot systems and a challenge in dynamically changing environments. By means of FPGAs it is possible to process time-critical and data-intensive tasks in robot systems efficiently. We have developed a parallel path planning algorithm which is based on Marching Pixels, an organic computing principle. The algorithm is optimized for FPGAbased processing. A parallel implementation approach together with an efficient buffering of the map data allowed us to obtain a processing rate suitable for realtime applications, even for higher resolutions. We achieved a processing rate of 80 maps per second for VGA resolution (640x480) on a midsize Virtex-5 FPGA.

show abstract

Section: Related Workmentioning

confidence: 99%

An Optimized FPGA Implementation for a Parallel Path Planning Algorithm Based on Marching Pixels

Schmidt

Fey

2010

2010 International Conference on Reconfigurable Computing and FPGAs

View full text Add to dashboard Cite

show abstract

“…), threat exposure, weather conditions, or any number of other factors. This set of cost values is known as the map cost (MC) array and is also of size m £ n [1,9,12].…”

Section: Preliminariesmentioning

confidence: 99%

“…The walking path, represented by the circled entries, corresponds to continually moving to the cell with the lowest BC value until the goal position is reached. There are a number of different ways to efficiently generate the BC array, any of which can be used [5,9]. Note that for the simple example shown in Fig.…”

Section: Preliminariesmentioning

confidence: 99%

“…Route planning is an important problem for a number of diverse applications including intelligent transportation systems, space applications, autonomous robotics, and military guidance and navigation systems [1,4,5,[9][10][11]. We address the routing problem in the context of route planning for aircraft (rotorcraft or fixed-wing, manned or unmanned), though our approach can be easily extended to the other applications outlined above.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust algorithm for real-time route planning

Szczerba

Galkowski

Glicktein³

et al. 2000

IEEE Trans. Aerosp. Electron. Syst.

285

120

View full text Add to dashboard Cite

“…This reduces the total number of messages needed. Finally, in [44], the costs of the straight path from goal to start is used for heuristic pruning, so more expensive paths are not explored further. On the CM-2, a 1024×1024 map is processed in 5.9 s. The processing time increase as O(l) with the path length l as long as there is a processor available for each map cell.…”

Section: Potential Field Approachesmentioning

confidence: 99%

A review of parallel processing approaches to motion planning

Henrich¹

Proceedings of IEEE International Conference on Robotics and Automation

View full text Add to dashboard Cite

. IntroductionOne of the many features needed to support the activities of autonomous systems is the ability to plan motion. Motion planning enables a robot to move in its environment securely and to accomplish a given task. In dynamic environments, the necessary adaptation of the robot motion is provided by closed control loops comprising sensing, planning, and acting, which have very short cycle times. One approach for realizing intelligent and reactive robotic systems is to integrate the planning algorithms into the control loop. Unfortunately, sound planning algorithms are complex and need long execution times. To still pursue this approach, a reduction in planning time is required. New parallel computing architectures with high computing power look promising.One could object that the motion planning problem in general can become very complex for increasing degrees of freedom (DOF) of the robot and is still intractable, even for a parallel computer. On the other hand, the aim of parallel processing is not to reduce the intractability of complex problems, but to reduce the solution time for given problems, or to increase their solution quality. Also, it is important to approximate the general problem by a simplified, but still realistic problem. This is independent or whether sequential of parallel processing is used.Thus, an important task is the parallelization of existing problem solutions in robotics so that they are suitable for highly computational parallelism. In several cases, fundamentally new algorithms have to be designed, so that a parallelization is feasible. The paper [28] reviews the research performed thus far in designing and implementing parallel algorithms for robotics. One of the key findings is that, in the subareas of manipulation and task planning, not much work has been published concerning parallel algorithms. Several parallel motion planing algorithms have been suggested however, which are reviewed in the rest of this paper. . Motion Planning ApproachesSequential motion planning approaches have been classified in: skeletons, cell decompositions, potential fields, and mathematical programming [16]. In the skeleton approach, the free configuration space (free C-space), i.e., the set of feasible motions, is retracted, reduced to, or mapped onto a network of one-dimensional (1D) lines. In the cell decomposition approach, the free C-space is decomposed into a set of simple cells, and the adjacency relationships among the cells are computed.To structure the parallel approaches, these keywords are modified by exchanging "skeletons" and "cell decompositions" with graph-based and grid-based approaches. The parallel ancillary algorithms, which are necessary to fulfill the planning task, are omitted due to space limitations. Graph-based ApproachesThe graph-based approaches include irregular skeletons and object-dependent cell decompositions. These approaches consist of two basic phases. The first phase is associated with the construction of a graph representing relations between free space. Aft...

show abstract

Highly paralellizable route planner based on cellular automata algorithms

Cited by 9 publications

References 12 publications

An Optimized FPGA Implementation for a Parallel Path Planning Algorithm Based on Marching Pixels

An Optimized FPGA Implementation for a Parallel Path Planning Algorithm Based on Marching Pixels

Robust algorithm for real-time route planning

A review of parallel processing approaches to motion planning

Contact Info

Product

Resources

About