Name
Title | ||
|---|---|---|
Regression-based supervised learning of autosteering of a road car featuring a delayed steering response |
Abstract | ||
|---|---|---|
In our work, we consider the area of intelligent road vehicles, especially the topic of automated vehicles. Our objective is to automatically develop—by applying a supervised learning approach—the steering of a realistically simulated car featuring both a steering delay and rate limit of the turning of its front wheels. Due to the adopted physical constraints, the typically used steering mechanisms (e.g., based on the servo-control model) result in a non-stable, oscillating behavior of the controlled car. The proposed approach of automated development of steering for such a realistically simulated car employs the perception–action relationship obtained from the sample runs of a simulated ideal car (featuring instant steering response) steered by the canonical servo-control model. Then, we proposed two approaches: The first one is based on our hypotheses that by offsetting the perceptions in the experimentally obtained perception–action relationship back in time to the value equal to the steering delay, we would be able to mimic the behavior of the car featuring steering delays and then used the modified perception–action relationship to train the proposed regression-based model. In the second approach, we first estimate the dynamics of the simulated car and then use this estimation for constructing the desired control function. The experimental results verify that the learned automated steering is able to control the car featuring steering delays of 100 and 200 ms in a much similar way as the servo-control steers an ideal car. Moreover, for delay of 400 ms, the steering learned via proposed approaches provides better quality of control than that obtained from one of the most versatile unsupervised machine learning approaches—genetic programming. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1007/s41060-018-0140-z | Journal of data science |
Keywords | Field | DocType |
Automatic steering, Delayed control feedback, Supervised learning, Regression-based Control | Regression,Simulation,Computer science,Control function,Quality of control,Supervised learning,Unsupervised learning,Rate limiting,Automatic steering | Journal |
Volume | Issue | ISSN |
7 | 2 | 2364-4168 |
Citations | PageRank | References |
1 | 0.40 | 9 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Vsevolod Nikulin | 1 | 5 | 1.59 |
| Albert Podusenko | 2 | 1 | 0.40 |
| Ivan Tanev | 3 | 278 | 46.51 |
| K. Shimohara | 4 | 23 | 5.56 |