Motion trajectory reconstruction and recognition with wireless devices are crucial in applications such as Virtual Reality (VR), sports performance monitoring, and intelligent medical systems. Traditional camera-based systems, though accurate, suffer from limited field of view, latency issues, and high costs. In contrast, wearable devices like Inertial Measurement Unit (IMU) sensors offer a compact and cost-effective alternative. Despite their advantages, low-cost IMUs face challenges like scaling errors, sensor axis misalignment, and non-zero biases. This paper proposes a novel Embedded System Platform (ESP) design for IMU-based trajectory reconstruction and recognition, incorporating dynamic calibration to address these challenges. Experimental results demonstrate the effectiveness of the proposed approach in enhancing motion trajectory reconstruction and recognition accuracy.