Tuning Feedback-Based Traffic Signal Controls
(2018)Department of Automatic Control
- Abstract
- A growing problem in many cities is that the traffic demand fluctuates heavily during the day and that it grows larger as the years go by. This calls for, and the technology to easier measure traffic enables, more efficient traffic light controls that can handle a greater traffic demand but also be adaptive. Centralised control of traffic light might, even though it can be both adaptive and efficient, be costly to install and expand due to the required communication between junctions. Decentralised control however, might be cheaper to install and maintain while still being adaptive and efficient. This thesis explores the Proportional Allocation controller with regards to its tuning parameter under conditions imposed by e.g., network size... (More)
- A growing problem in many cities is that the traffic demand fluctuates heavily during the day and that it grows larger as the years go by. This calls for, and the technology to easier measure traffic enables, more efficient traffic light controls that can handle a greater traffic demand but also be adaptive. Centralised control of traffic light might, even though it can be both adaptive and efficient, be costly to install and expand due to the required communication between junctions. Decentralised control however, might be cheaper to install and maintain while still being adaptive and efficient. This thesis explores the Proportional Allocation controller with regards to its tuning parameter under conditions imposed by e.g., network size and topology, and queue sensor length. Two aspects of its performance were measured:
optimality and fairness. In addition to the original controller two modifications are proposed and evaluated. It was found that when simulating, smaller networks could be used to approximate bigger ones, and that the queue sensor length is of critical importance not to be too small. The proposed modifications were found to perform similarly to the original one even though it in principal should not. The results herein should give a good starting point of the optimal value of the tuning parameter in a real-world implementation. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8960386
- author
- Paulsson, Simon
- supervisor
-
- Gustav Nilsson LU
- Giacomo Como LU
- Anton Cervin LU
- organization
- year
- 2018
- type
- H3 - Professional qualifications (4 Years - )
- subject
- report number
- TFRT-6067
- ISSN
- 0280-5316
- language
- English
- id
- 8960386
- date added to LUP
- 2018-10-29 10:46:58
- date last changed
- 2018-10-29 10:46:58
@misc{8960386, abstract = {{A growing problem in many cities is that the traffic demand fluctuates heavily during the day and that it grows larger as the years go by. This calls for, and the technology to easier measure traffic enables, more efficient traffic light controls that can handle a greater traffic demand but also be adaptive. Centralised control of traffic light might, even though it can be both adaptive and efficient, be costly to install and expand due to the required communication between junctions. Decentralised control however, might be cheaper to install and maintain while still being adaptive and efficient. This thesis explores the Proportional Allocation controller with regards to its tuning parameter under conditions imposed by e.g., network size and topology, and queue sensor length. Two aspects of its performance were measured: optimality and fairness. In addition to the original controller two modifications are proposed and evaluated. It was found that when simulating, smaller networks could be used to approximate bigger ones, and that the queue sensor length is of critical importance not to be too small. The proposed modifications were found to perform similarly to the original one even though it in principal should not. The results herein should give a good starting point of the optimal value of the tuning parameter in a real-world implementation.}}, author = {{Paulsson, Simon}}, issn = {{0280-5316}}, language = {{eng}}, note = {{Student Paper}}, title = {{Tuning Feedback-Based Traffic Signal Controls}}, year = {{2018}}, }