This paper proposes to apply the method with an additional noisy layer to the structure of the graph neural network for reinforcement learning in automated design technology for information and communication systems. The automated design technology has an elementary problem of huge learning time caused by overestimating a specific configuration because of hardly ever rewards despite huge exploration space with a vast combination of selections, arrangements, and connections. The parametric noise applied to the network is learned with gradient descent and the remaining network weights to reduce this harmful overestimation during learning and increase the design exploration efficiency. The evaluation result showed that using the proposed algorithm for our automated design technology in development could shorten 15% of the episodes needed for learning to converge.
With the spread of Internet of Things (IoT) technology, the requirements for antennas mounted on electrical devices have increased. These antennas are required to have a small size and thin profile, and to be unaffected by metals. A metasurface-inspired antenna chip developed by the KIT EOE Laboratory (MACKEY) is proposed as an antenna that satisfies these conditions. Among IoT devices, smartphones and smartwatches use circularly polarized waves to accommodate polarization at various angles. Therefore, MACKEY, which radiates circularly polarized waves, has also been proposed. In this paper, we propose a one feeding-point model with broadband axial ratio characteristics.
In this paper, a low loss micro-strip line (MSL) on thin flexible substrate was realized using defected ground structure (DGS). DGS allows thinner MSL than conventional MSL maintaining insertion loss and far-end crosstalk. MSL with DGS on the flexible substrate with a thickness 25 µm was realized the same insertion loss and far-end crosstalk in comparison with conventional MSL on the substrate with a thickness 50 µm while analyzing by 3D electromagnetic simulation. The MSL with DGS was fabricated and evaluated to confirm the simulated results. Excellent agreement with the measurement results was demonstrated over the broad bandwidth of 1-40 GHz.