Title:
5-axis coverage path planning with deep reinforcement learning and fast parallel collision detection
5-axis coverage path planning with deep reinforcement learning and fast parallel collision detection
| dc.contributor.advisor | Vuduc, Richard | |
| dc.contributor.author | Chen, Xin | |
| dc.contributor.committeeMember | Kurfess, Thomas | |
| dc.contributor.committeeMember | Catalyurek, Umit | |
| dc.contributor.committeeMember | Young, Jeff | |
| dc.contributor.committeeMember | Tucker, Thomas | |
| dc.contributor.department | Computer Science | |
| dc.date.accessioned | 2020-05-20T17:02:09Z | |
| dc.date.available | 2020-05-20T17:02:09Z | |
| dc.date.created | 2020-05 | |
| dc.date.issued | 2020-04-24 | |
| dc.date.submitted | May 2020 | |
| dc.date.updated | 2020-05-20T17:02:09Z | |
| dc.description.abstract | 5-axis machining is a strategy that allows computer numerical control (CNC) move an object or cutting tool along five different axes (X, Y, Z and two additional rotary axes) simultaneously. This provides infinite possibilities of machining very complex objects, which is why 5-axis machining gets more and more popular. This thesis focuses on a path planning problem that arises in 5-axis machining applications: how to construct a tool path that covers the surface of a 3D object, produces a short milling time, and is collision-free. This thesis proposes a general path planning framework with a fast collision detection algorithm to generate an efficient 5-axis path. We first present a unifying, general and adaptive framework with deep reinforcement learning, called adaptive deep path (AD Path), to generate an efficient path for covering an arbitrary 2D environment. The key idea of this algorithm is a new graph model based on boustrophedon cellular decomposition (BCD), which is a method of transforming a space into cell regions with morse decomposition. This graph model can easily reflect the physical distance in the graph, and evaluate the cost of an arbitrary path. We show that when applied to deep reinforcement learning, AD Path can efficiently reduce the path length and the number of corners adaptively. Second, this thesis presents a fast parallel collision detection algorithm, named aggressive inaccessible cone angle (AICA) for CNC milling applications. The key idea of our proposed method is the concept of inaccessible cone angle (ICA), which is a new geometric abstraction for collision detection tests, and its effective use, including memoization to remove redundant work and increasing parallelization. We have prototyped our AICA algorithm within a real CNC milling tool, SculptPrint. Experimental results on 4 CAD benchmarks demonstrate that AICA is up to 23 times faster than the approach of the traditional checking. Third, this thesis proposes a new 5-axis coverage path planning algorithm, called max orientation coverage, considering both the trajectory of the cutting tool end in 3-axis, and the orientations of the tool as the other 2 rotatory axes. This algorithm aims at reducing the machining time, by designing an efficient 5-axis path to reduce the number of tool reorientations and the number of tool retractions (pulling the tool back and in) as a constraint of being collision-free. Our proposed method employs a two-step optimization. We validate our method using four CAD benchmark objects against a previously proposed random sampling-based coverage algorithm. On average, our method improves the path efficiency by 29.7%. | |
| dc.description.degree | Ph.D. | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.uri | http://hdl.handle.net/1853/62825 | |
| dc.language.iso | en_US | |
| dc.publisher | Georgia Institute of Technology | |
| dc.subject | Coverage path planning | |
| dc.subject | Collision detection | |
| dc.subject | Deep reinforcement learning | |
| dc.title | 5-axis coverage path planning with deep reinforcement learning and fast parallel collision detection | |
| dc.type | Text | |
| dc.type.genre | Dissertation | |
| dspace.entity.type | Publication | |
| local.contributor.advisor | Vuduc, Richard | |
| local.contributor.corporatename | College of Computing | |
| local.contributor.corporatename | School of Computer Science | |
| relation.isAdvisorOfPublication | e9a36794-e148-4304-8933-6ae0449c21d2 | |
| relation.isOrgUnitOfPublication | c8892b3c-8db6-4b7b-a33a-1b67f7db2021 | |
| relation.isOrgUnitOfPublication | 6b42174a-e0e1-40e3-a581-47bed0470a1e | |
| thesis.degree.level | Doctoral |