Integrating Task and Motion Planning
- Authors
- C?rdenas Monar, Jorge Luis
- Format
- MasterThesis
- Status
- publishedVersion
- Description
Integrating task and motion planning is a computational dif?cult problem which has a many of interesting applications in the real world. There are several ways to integrate task and motion planning, but usually authors take the approach to use motion planners as the core of their development. This approach have been deeply studied over the years and achieved useful results [Cambon et al. 2009; Plaku and Hager 2010]. However, the problem is far from being solved. The different characteristics of the task planner (discrete variables, ?nite search space) and motion planners (continuous variables, in?nite con?guration space) makes them hard to combine and even more complicated to get a sound and complete solution with great performance. The recent work by [Ivankovic et al. 2014] extends classical planning to integrate constraints given to an external solver. Here we adapt this work to integrate task planning with the motion planner OpenRAVE, which plays the role of the external solver. This Thesis attempts to correctly formalize, implement, and simulate the combined task and motion planning problem.
- Publication Year
- 2015
- Language
- eng
- Topic
- COMPUTACI?N
INTELIGENCIA ARTIFICIAL
APLICACI?N INFORM?TICA
TECNOLOG?A DE LA COMUNICACI?N
- Repository
- Repositorio SENESCYT
- Rights
- openAccess
- License