Adaptive pursuit learning method to mitigate small-cell interference through directionality
Overview
This technology is a learning protocol for distributed antenna state selection in directional cognitive small-cell networks. The protocol addresses interference and performance improvement in wireless data transmission systems, specifically focusing on small-cell deployments with directional antennas. It formulates the antenna state selection as a nonstationary multi-armed bandit problem and proposes a solution using the adaptive pursuit method from reinforcement learning. The system leverages a practical implementation, referred to as WARP-TDMAC, which integrates electronically reconfigurable antennas into small-cell networks, enabling synchronized directional transmission.
Market Applications
- Small-Cell Deployments: The protocol is applicable to dense small-cell networks, including 5G deployments, to improve network capacity and reduce interference
- Wireless Communication Systems: It can be applied to various wireless communication systems seeking to enhance link performance and network throughput
- Machine Learning in Networking: The use of machine learning and adaptive algorithms for antenna state selection has broader applications in network optimization
Key Advantages
- Improved Link Performance: The protocol uses machine learning-based distributed antenna state selection to enhance individual transmission links' performance by avoiding interference and adapting to changing environments
- Total System Throughput: The patent introduces synchronous directional transmission across the network, improving total system throughput by enabling spatial reuse and efficient channel access
- Adaptive Pursuit Algorithm: The use of the adaptive pursuit algorithm in antenna state selection allows for dynamic adaptation to non-stationary environments, making it suitable for real-world scenarios
Problems Solved
- Interference Mitigation: The patent addresses the challenge of interference in dense small-cell deployments by using directional antennas and machine learning for optimal antenna state selection
- Integration of Directional Antennas: It provides a solution for integrating directional antennas into small-cell networks, overcoming protocol overhead and adaptation difficulties
- Efficient Network Synchronization: The patent proposes a hybrid synchronization mechanism and real-time scheduling for achieving synchronization in a multi-link wireless network
Intellectual Property
United States Issued Patent- Adaptive pursuit learning method to mitigate small-cell interference through directionality
References
A. Paatelma, D. H. Nguyen, H. Saarnisaari, N. Kandasamy, and K. R. Dandekar, “Reinforcement Learning System to Mitigate Small-Cell Interference Through Directionality,” in Proc. IEEE Intl. Symp. on Personal, Indoor, and Mobile Radio Communications (PIMRC), 2017.
D. H. Nguyen, A. Paatelma, H. Saarnisaari, N. Kandasamy, and K. R. Dandekar, “Demo: Enhancing Indoor Spatial Reuse through Adaptive Antenna Beamsteering,” in Proc. ACM Intl. Workshop on Wireless Network Testbeds, Experimental Eval., and Characterization (WiNTECH), 2016.
Contact Information
For Intellectual Property and Licensing inquiries
Inventor information