Fused smart sensor network for multi-axis forward kinematics estimation in industrial robots

Area of Science:

• Robotics
• Sensor Networks
• Control Systems

Background:

• Industrial robots require accurate position and orientation sensing (forward kinematics) for optimal performance.
• Traditional optical encoders fail to detect mechanical deformations, compromising robot accuracy.
• Existing sensor fusion methods often incur high computational costs, hindering real-time applications.

Purpose of the Study:

• To propose a novel fused smart sensor network for estimating industrial robot forward kinematics.
• To enable online measurement of joint angular position and real-time forward kinematics estimation.
• To enhance robot accuracy by detecting and compensating for mechanical deformations.

Main Methods:

• A smart sensor network integrating optical encoders and 3-axis accelerometers.
• Utilizing Kalman filters for sensor data fusion and noise reduction.
• Implementing the system on a field-programmable gate array (FPGA) for parallel processing and online computation.

Main Results:

• The developed system effectively fuses data from multiple sensors.
• Online estimation of joint angular position and forward kinematics was achieved.
• The smart sensor network demonstrated robust performance in real-operation scenarios on a 6-DOF robot.

Conclusions:

• The proposed fused smart sensor network offers an efficient solution for accurate industrial robot forward kinematics estimation.
• FPGA implementation enables real-time processing, overcoming limitations of high-computational load methods.
• This approach enhances robot accuracy and reliability in industrial applications.