深度强化学习技术
Search documents
嘉陵江实验室激活制造升级新动能
Ren Min Ri Bao Hai Wai Ban· 2025-12-03 01:04
Core Insights - The Jialing River Laboratory is advancing manufacturing upgrades by focusing on intelligent robotics and smart equipment, utilizing AI and deep reinforcement learning for autonomous four-legged robots capable of navigating complex terrains [1][2] - The laboratory aims to bridge the gap between research and production, aligning with national strategies to enhance the overall effectiveness of the innovation system [2] - The laboratory has been recognized as a key center for intellectual property operations, addressing the issue of dormant patents and facilitating technology transfer to industry [2] Group 1 - The Jialing River Laboratory has developed autonomous four-legged robots that can adapt to various terrains, enhancing their application in underground search and rescue and logistics [1] - The laboratory is part of a broader initiative to create a research platform focused on smart equipment, including intelligent robots and smart medical devices, with over 200 researchers involved [1][2] - The laboratory's technology transfer team is crucial in ensuring that research outcomes are effectively commercialized, emphasizing the importance of integrating research with industry needs [2] Group 2 - The laboratory will implement a "challenge and lead" mechanism to drive innovation in key areas, specifically targeting the development of intelligent robots for complex underground environments [3] - The focus is on producing market-ready products and fostering new industries, moving beyond traditional academic outputs like papers and patents [3] - The laboratory has been selected as one of the first national intellectual property operation centers, collaborating with leading companies to create a model that connects research with practical applications [2][3]
国网兰州供电公司研发全省首个通信网故障监测智能体
Zhong Guo Neng Yuan Wang· 2025-11-07 10:02
Core Insights - The State Grid Lanzhou Power Supply Company, in collaboration with the State Grid Gansu Electric Power Company Dispatch Center, has developed an AI-based communication network fault monitoring agent, which is currently in trial operation [1][2]. Group 1: AI Implementation and Technology - The communication network, described as the "nervous system" of the power grid, is becoming increasingly complex and critical, prompting the need for enhanced monitoring and management solutions [1]. - The intelligent agent utilizes deep reinforcement learning technology, incorporating power communication regulations, technical procedures, network topology information, and over a hundred diversion strategies and typical cases to create a professional knowledge graph [2]. Group 2: Operational Efficiency and Performance Metrics - Since its trial operation, the intelligent monitoring agent has been monitoring an average of 456 communication network elements daily, providing six fault handling suggestions with an accuracy rate of 90% [2]. - The agent has simulated five optical cable maintenance operations, assessing the impact on a total of 164 power grid services, and has automatically identified risks associated with large power grid operations, achieving a risk list accuracy of 97% [2]. Group 3: Future Developments - The company plans to continuously iterate and optimize the intelligent agent, extending its capabilities to include precise analysis of communication network security alerts and intelligent handling functions, thereby enhancing communication dispatch efficiency and safety management levels [2].