Detection And Tracking Of Microrobotic Multi-manipulators For Cellular Scaffold Assembly

Micro-nano manipulation system refers to robotic system which can be used to do high-precision operation tasks under micron, submicron and nano scale. It provides new opportunities and techniques for the technical development of electronics, information, materials, advanced manufacturing and biomedicine. The loss of organ or tissue is currently one of the most frequent, devastating and costly problems in human health care. Micro-nano manipulation also provides new concept significance in tissue engineering and demonstrates a paradigm for the fussion of micro-nano robotics and biomedical science through multi-scaled motion and automated cooperation in the microrobotic system. As a main control method in micro-nano manipulation system, micro visual servo becomes a hot issue in the field of micro and nano operation which has a significant research value. The main research issues and results are summarized as follows.Firstly, according to the bottom-up assembly principle of 3D cellular microstructure, we develop an on-chip fabrication method of 2D cellular modules as the assembly units. Through the UV-exposure of crosslinkable hydrogel which mixed with cells in the microfluidic channel, the 2D cellular micromodules are made for the following assembly. The micromodules are assembled to a tubular structure through repeating a newly designed single-step pick-up motion.Secondly, based on the phenomenon that image is clear when it is on the focused plane and blur on defocused plane, we propose an automatic focusing method for focused image acquisition before micro-manipulation. The relative depth information of the micromodules and the tips of micromanipulator can be also obtained.Thirdly, according to the contour features of fabricated cellular micromodules, we proposed a target detection method to recognize the micromodules. This method reconstructs the target contour with partial contour information based on contour area method when the target occludes with background. Experimental results show that the method performance good with high detection accuracy.Finally, because of the low contrast between micro-manipulator and cellular micromodules in the micro-scaled image, it is difficult to track multi tips of micro-manipulators. We combine the level set method and particle filtering to track the tips with similar features during the pick-up progress. Experimental results shows that this algorithm can effectively solve the tracking loss of multi target with similar features under microscale.