Home>Categories Detail

Categories

Robot and Automation

2.1.3 Robot and automation

The present research of robot and automation covers the principles of dynamic walk, leg of robot, exoskeleton and artificial limb. Associate Prof. Chenglong Fu and Prof. Caihua Xiong charge research in this topic.

2.1.3.1 Automatic & dynamic artificial limb of robot


The metabolism of people wearing conventional artificial limb is 30% greater than that of normal people.


 The conventional dynamic artificial limb cannot be comparable with healthy limb.
 

The dynamic artificial limb is able to simulate the biomechanical property of knee and ankle.

 At the present, the majority of commercialized artificial limbs are passive type, which costs 60% more metabolism energy than that of normal people. Moreover, the hip torque generated from artificial limb is three times of that from normal people. We designed a dynamic artificial limb including the ankle and knee joints to simulate the biomechanical property of human walk and greatly reduce the metabolism energy cost, which facilitates complex actions, such as walking on heavy load, walking up and down stairs and standing up.

2.1.3.2 Force assisted exoskeleton




We proposed an exoskeleton that recycles the energy from knee and ankle to force the ankle joint acting on the ground. Specifically, the negative power made by the end stage of knee joint motion and initial stage of ankle motion, is recycled from a controllable dual surface wheel to torsion spring, which further support the action of ankle joint acting on the ground. Experiment shows that the proposed exoskeleton can greatly reduce the motion of hamstrings and gastrocnemius muscles.

2.1.3.3 Biped robot with dynamic walk


Biped robot with dynamic walk (THR-I) is realized by retraction of swing leg and sensory reflex control, which can be dynamic high speed walk and running at a speed of two times of the length of leg every second. The maximum stride is 0.56 times of the length of leg with a minimum period of 0.2 second.



The bionic robot THR-II is powered by human muscle. The control system is based on the combination of global feedback and sensory reflex. The foothold is dynamically updated with a steady recovery of 100N×0.2s to turbulence. A book focusing on the bionic robot has been published.

2.1.3.4  Legged robot and quadruped robot




The legged robot with energy conservation driven by jumping is designed. The effect of driven form and action time to the energy of jumping property is studied to elucidate the strategy of driving force and damping energy.



The control of quadruped robot is proposed by establishing a relationship between high level walking task bounded by contacting force the foot and the mapping of joint torsion.

2.1.3.5 The saving mechanism of carrying-pole



The project combines theoretical study and experiment to studying the terms of loading boundary to mechanical property during walking and the energy efficiency. The saving mechanism of carrying-pole, which is a transportation tool in ancient China. The wisdom from ancient China is casted to the world. Meanwhile, it provides novel method guiding the boundary terms of heavy load to humanoid robot and exoskeleton.

2.1.3.6 Modeling and identification of natural gestures




The existing modeling and identification of natural gesture are based on the trajectory of palm motion, which does not consist with the habit of natural gesture and needs lots of samples to training the model. Moreover, it is hard to add and modify gesture. The current project expresses the gesture of joint motion, action type and arm shape one by one. By different forms of the judgement combination, it can be naturally and efficiently model large amount of gestures. Additionally, it can process the complex gestures that cannot be dealt by the previous methods. When it comes to identification, the proposed method only requires simply judging the similarity between the current gesture and existing gesture, to ensure the high accuracy and reduced computational load.