| Sumario: | The poses of robotic endoscopic capsules are indispensable for further follow-up examinations, potential targeted drug delivery, and closed-loop controlling of active locomotion. A novel tracking method using the multiple magnetic excitations with frequency division has been investigated. The multiple excitation coils can simultaneously work at different frequency to improve real-time tracking. A novel model between the magnetic flux density and the capsule’s pose has been derived, which shows a nonlinear equation group with multiple local extremum. Then, a Back-Propagation (BP) neural network algorithm combined with the mother wavelet is investigated to solve the pose. To reduce the volume and power consumption, the wireless magnetic sensing module uses digital signal processing as the core framework, which is beneficial to be miniaturized to integrate with the capsule. The functional prototype of the tracking system has been developed, which consists of a wireless magnetic sensing module mounted in the capsule, a magnetic excitation module with frequency division, a wireless receiver and data interface, an excitation coil array and a platform for pose solving. The experimental results show that the mean errors are 0.0098 m in x-component, 0.0122 m in y-component, 0.0077 m in z-component, 0.187 rad in α-component and 0.161 rad in β-component, respectively. The real-time performance of the tracking system is improved.
|
|---|