TinyLight: Adaptive Traffic Signal Control on Devices with Extremely Limited Resources
Authors: Dong Xing, Qian Zheng, Qianhui Liu, Gang Pan
IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate Tiny Light on multiple road networks with real-world traffic demands. Experiments show that even with extremely limited resources, Tiny Light still achieves competitive performance. |
| Researcher Affiliation | Academia | 1College of Computer Science and Technology, Zhejiang University, Hangzhou, China 2Department of Electrical and Computer Engineering, National University of Singapore, Singapore {dongxing, qianzheng}@zju.edu.cn, qhliu@nus.edu.sg, gpan@zju.edu.cn |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The source code and appendix of this work can be found at https://bit.ly/38h H8t8. |
| Open Datasets | Yes | We evaluate our work on six road networks with real-world traffic flows, which are all publicly available.4 Figure 3 presents a top view of these road networks... |
| Dataset Splits | No | The paper does not explicitly provide training/validation/test dataset splits with specific percentages or counts. It mentions creating nine traffic flows by shifting initial appearance times, but this is not a train/validation/test split. |
| Hardware Specification | Yes | This enables Tiny Light to work on a standalone microcontroller with merely 2KB RAM and 32KB ROM... As a demonstrative prototype of our model s deployment, we implement Tiny Light on a standalone ATmega328P an MCU with merely 2KB RAM and 32KB ROM. This MCU is readily available on the market and costs less than $5, making our work applicable in scenarios with low budgets. |
| Software Dependencies | No | The paper states it avoids dependencies on certain third-party libraries (Tensor Flow Lite, Py Torch) but does not list specific version numbers for the software components it does use for its experiments or development, other than mentioning 'City Flow' without a version. |
| Experiment Setup | No | The paper states that "Appendix B provides more detailed settings for training." While it mentions the optimization method (dual gradient descent) and reward function definition, it does not provide concrete hyperparameter values or comprehensive training configurations in the main text. |