Search

In this paper, we propose the through silicon via (TSV) array design optimization method using deep reinforcement learning (DRL) framework. The agent trained through the proposed method can provide an optimal TSV array that minimizes far-end crosstalk (FEXT) in one single step. We define the state, action, and reward that are elements of the Markov Decision Process (MDP) for optimizing the TSV array considering FEXT and train a deep q network (DQN) agent. For verification, we applied the proposed method to a 3 by 3 through silicon via array at stacked DRAM of High Bandwidth Memory (HBM). The network converged well, and as the result, the proposed method provided the optimal design that satisfies the target FEXT in which 3 dB lower than the initial design.

In this paper, we propose artificial intelligent (AI)router, a reinforcement learning (RL)-based auto-router considering signal integrity (SI), for the first time. Our algorithm has two main stages. At first, we design the transformer-based novel neural architecture considering the keep-out region, crosstalk region, and the number of vias for SI optimization. Then, the designed neural network is optimized by the policy gradient, one of the RL algorithms. Compared with the conventional maze routers, the A* algorithm, and the lee algorithm, it is verified that our AI-router outperforms the algorithms in terms of wire-length and crosstalk in a specific test case. Furthermore, it is shown that AI-router successfully performs multi-layer routing which is not feasible with conventional maze routers.

Recently, vehicle driving devices to support infotainment services are being designed for driver’s convenience. Especially, the automobile industry is developing vehicles that use a joy stick instead of a steering wheel from the concept car. The greatest strength of using a joy stick is that the driver feels less workload and fatigue than when using a steering wheel. However, this field still needs more research and experiments for more accurate results. Therefore, this study evaluated the workload when driving with different devices through a subjective questionnaire and measurement of physiological signals. The reason for using not only a survey but also physiological signals is to support the results of the subjective questionnaire, which are not enough to yield accurate results. There are two different kinds of methods to carry out this research: SWAT (subjective workload assessment technique) for the survey and Biopac equipment for the measurement of physiological signals. A thermal infrared camera is used for capturing the driver workload by measuring the facial skin temperature. Furthermore, a previously established driving simulator, GPS (global positioning system), and the virtual Kyeong-Bu highway are used for the experiment. As a result of the experiment with 13 subjects following both methods (the surveys and the measurement of physiological data), it is confirmed that using a drive stick device causes less workload and fatigue to the drivers than using a steering wheel. In addition, the statistical significance of the results is verified through the SPSS (statistical package for the social sciences) statistical analysis program.