Design considerations of reinforcement learning power controllers in Wireless Body Area Networks
- Publication Type:
- Conference Proceeding
- Citation:
- IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, PIMRC, 2012, pp. 2030 - 2036
- Issue Date:
- 2012-12-01
Closed Access
| Filename | Description | Size | |||
|---|---|---|---|---|---|
| 06362688.pdf | Published version | 531.55 kB |
Copyright Clearance Process
- Recently Added
- In Progress
- Closed Access
This item is closed access and not available.
A Wireless Body Area Network (WBAN) comprises a number of tiny devices implanted in/on the body that sample physiological signals of the human body and send them to a coordinator node for medical or other purposes. As these miniature devices run on built-in batteries, energy is the most valuable resource in WBANs. This makes signal interference between neighboring WBANs a serious threat because it causes energy waste in these systems. To mitigate this internetwork interference, we propose a dynamic power control mechanism in WBANs which employs reinforcement learning (RL) to learn from experience and improve its performance. This paper presents guidelines in designing efficient RL power controllers in WBANs and provides an analysis of the effect of the reward function, discount factor, learning rate and eligibility trace parameter where the main performance criteria used are convergence and solution optimality in terms of throughput and energy consumption per bit. © 2012 IEEE.
Please use this identifier to cite or link to this item: