Recently, Wenxing Li, a Ph.D. student at the State Key Lab of Processors (SKLP), Institute of Computing Technology, Chinese Academy of Sciences (supervised by Professor Huawei Li), was awarded the sole Best Paper Award of the 32nd IEEE ATS, at the opening ceremony of ATS 2024 (held in Ahmedabad, India), for the paper titled Intelligent Automatic Test Pattern Generation for Digital Circuits Based on Reinforcement Learning, in which he is the first author.
Automatic test pattern generation (ATPG) is a critical technology in integrated circuit testing. It searches for effective test vectors to detect all possible faults in the digital circuit under test, thereby ensuring high fault coverage and ensuring chip quality after testing. In the branch-and-bound based test pattern search process, invalid backtrace path selection will cause significant increase of number of backtracks during the test generation for very-large-scale integration (VLSI) circuits, which affects the performance of ATPG. The award-winning paper proposes an ATPG method guided by reinforcement learning (RL), which defines the state space, action space, state transition and reward function consistent with the RL tasks in the circuit environment. The solution effectively guides the agent to learn an efficient backtrace strategy. Experimental results show that the proposed RL-ATPG can achieve better performance in comparison with the previous ATPG methods using traditional heuristic strategies or strategies based on deep learning.
The IEEE Asian Test Symposium (ATS), founded in 1992, has been sponsored by the IEEE Computer Society (CS) Test Technology Technical Community (TTTC) since its foundation, and co-sponsored by the IEEE Council on Electronic Design Automation (CEDA) in recent years. It is listed as a Class-C conference in the CCF (China Computer Federation) recommendation list. Among the list of accepted papers in each year, only one Best Paper Award was selected by the ATS Best Paper Award Selection Committee.
downloadFile