This paper introduces a skeleton calibration method for orientation measurement sensors based motion capture systems. In this method, the orientation sensors are used to measure the posture of the limbs. A template is used to register the end-effectors' postures (feet, hands for example). Through capturing the limb postures while the end-effectors match with the pre-defined postures on the template, a linear equation system of the skeleton dimensions can be generated based on the human kinematics. The limb dimension parameters can be optimized based on that. The identifiablility of this skeleton dimension is discussed. The symmetric property of the skeleton is also taken into consideration. To demonstrate the method, IMU sensors and a footprint template are used to calibrate the lower limb dimension. Results show the absolute dimension errors can be controlled within centimeter level for this human lower limb calibration. Compared with existing methods, this template based method is a quick and self-contained method which does not need extra measurement devices, and the skeleton model does not have asymmetric problems. Since the inertial MoCap systems are widely used nowadays, this method is useful to generate an accurate skeleton model for precise behavior presentation.

Robotic compliance control is critical for delicate tasks such as electronic connector assembly, where precise force regulation and adaptability are paramount. However, traditional methods often struggle with modeling inaccuracies and sensor noise. Inspired by human adaptability in complex assembly operations, we present RoboMT, a novel framework that integrates a Mamba algorithm with a Transformer architecture to achieve human-like compliance control. By leveraging a bilateral teleoperation platform, we collect extensive real-time force/torque and motion data to form a comprehensive dataset for training. Furthermore, RoboMT incorporates an Adaptive Action Chunk module and a Temporal Fusion module to ensure smooth and robust action prediction. Experimental results across four electronic assembly tasks show that RoboMT achieves superior success rates (62–98%) over baselines (29–98%), while maintaining stable force regulation around 2.5 N, closely resembling human performance. During task transitions, RoboMT quickly stabilizes at 5 N with minimal overshoot, avoiding the large force spikes (over 24 N) seen in baselines. Additionally, RoboMT maintains an average inference speed of 55 ms per batch, balancing real-time responsiveness and control robustness. Overall, RoboMT presents a compelling pathway toward error-minimized, human-level compliance control, and generalization for real-world robotic assembly, setting a new benchmark for precision, adaptability, and robustness in robotic assembly.