Abstract
In a previous paper, we proposed a solution to motion planning of a mobile robot. In our approach, we formulated the problem as a discrete optimization problem at each time step. To solve the optimization problem, we used an objective function consisting of a goal term, a smoothness term and a collision term. In this paper, we propose a theoretical method using reinforcement learning for adjusting weight parameters in the objective functions. However, the conventional Q-learning method cannot be applied to a non-Markov decision process, which is caused by the smoothness term. Thus, we applied William's learning algorithm, episodic REINFORCE, to derive a learning rule for the weight parameters. This maximizes a value function stochastically. We verified the learning rule by some experiments.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the IEEE International Symposium on Assembly and Task Planning |
| Pages | 1-6 |
| Number of pages | 6 |
| Publication status | Published - 2001 |
| Externally published | Yes |
| Event | 2001 IEEE International Symposium on Assembly and Task Planning (ISATP2001) - Fukuoka Duration: 2001 May 28 → 2001 May 29 |
Other
| Other | 2001 IEEE International Symposium on Assembly and Task Planning (ISATP2001) |
|---|---|
| City | Fukuoka |
| Period | 01/5/28 → 01/5/29 |
Fingerprint
ASJC Scopus subject areas
- Engineering(all)
Cite this
Motion planning of a mobile robot as a discrete optimization problem. / Igarashi, Harukazu.
Proceedings of the IEEE International Symposium on Assembly and Task Planning. 2001. p. 1-6.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Motion planning of a mobile robot as a discrete optimization problem
AU - Igarashi, Harukazu
PY - 2001
Y1 - 2001
N2 - In a previous paper, we proposed a solution to motion planning of a mobile robot. In our approach, we formulated the problem as a discrete optimization problem at each time step. To solve the optimization problem, we used an objective function consisting of a goal term, a smoothness term and a collision term. In this paper, we propose a theoretical method using reinforcement learning for adjusting weight parameters in the objective functions. However, the conventional Q-learning method cannot be applied to a non-Markov decision process, which is caused by the smoothness term. Thus, we applied William's learning algorithm, episodic REINFORCE, to derive a learning rule for the weight parameters. This maximizes a value function stochastically. We verified the learning rule by some experiments.
AB - In a previous paper, we proposed a solution to motion planning of a mobile robot. In our approach, we formulated the problem as a discrete optimization problem at each time step. To solve the optimization problem, we used an objective function consisting of a goal term, a smoothness term and a collision term. In this paper, we propose a theoretical method using reinforcement learning for adjusting weight parameters in the objective functions. However, the conventional Q-learning method cannot be applied to a non-Markov decision process, which is caused by the smoothness term. Thus, we applied William's learning algorithm, episodic REINFORCE, to derive a learning rule for the weight parameters. This maximizes a value function stochastically. We verified the learning rule by some experiments.
UR - http://www.scopus.com/inward/record.url?scp=0034826829&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0034826829&partnerID=8YFLogxK
M3 - Conference contribution
SP - 1
EP - 6
BT - Proceedings of the IEEE International Symposium on Assembly and Task Planning
ER -