No abstract

In this paper, we analyze the precision of the tilting angles transferred from a 3-axis accelerometer with the previously proposed CORDIC-based tilting angle identification algorithm. We found out two issues of the original 2D-CORDIC operations which would increase the errors of the calculated angles. Hence, we propose the solutions to further enhance the precision of tilting angles calculated in the algorithm. We also evaluate the performance of our algorithm on 5 different platforms with different processor cores ranging from low-end micro-controllers to high-end desktop CPU. The results indicate that the algorithm is extremely suitable for implementation on low cost and low power micro-controllers to perform the real-time and in-line tilting angle computations for its low computation complexity and low memory requirements. As the results, precise tilting angles can be obtained in realtime with a low cost embedded system to be applied in various applications including medical care systems.

No abstract

No abstract

Since the power integrity problem has become one of the critical issues that limit the design performance, it is essential to obtain the supply current waveforms at early design stages to make efficient reduction of the supply noise. Therefore, a high-level macro current model is proposed in [11] for logic blocks to provide quick current waveform estimation at RTL. However, due to the different arrival time of internal signals, it is not easy to accurately model the supply current of the whole logic blocks as some fixed templates. Therefore, a levelized high-level current model is proposed in this work for macro blocks. By grouping those gates with similar arrival time as a super-gate, the current waveform of this super-gate can be recorded easily with less error. Then, combining all triangles of every super-gate in time obtains a more accurate supply current waveform, especially for multi-peak cases. In order to consider the different arrival time of different input patterns, a dynamic levelization algorithm is also proposed. As shown in the experimental results, the average peak error of the proposed levelized current model is improved by almost 23% compared to the conventional single-stage current model.