A type of wavelet neural network, in which the scale function isadopted only, is proposed in this paper for non-linear dynamicprocess modelling. Its network size is decreased significantly andthe weight coefficients c...A type of wavelet neural network, in which the scale function isadopted only, is proposed in this paper for non-linear dynamicprocess modelling. Its network size is decreased significantly andthe weight coefficients can be estimated by a linear algorithm. Thewavelet neural network holds some advantages superior to other typesof neural networks. First, its network structure is easy to specifybased on its theoretical analysis and intuition. Secondly, networktraining does not rely on stochastic gradient type techniques andavoids the problem of poor convergence or undesirable local minima.展开更多
The bulk TC17was subjected to the high energy shot peening(HESP)at the air pressures ranging from0.35to0.55MPa and processing durations ranging from15to60min.The microhardness(HV0.02)from topmost surface to matrix of ...The bulk TC17was subjected to the high energy shot peening(HESP)at the air pressures ranging from0.35to0.55MPa and processing durations ranging from15to60min.The microhardness(HV0.02)from topmost surface to matrix of the HESP processed TC17was measured,which generally decreases with the increase of depth from topmost surface to matrix and presents different variation with air pressure and processing duration at different depths.A fuzzy neural network(FNN)model was established to predict the surface layer microhardness of the HESP processed TC17,where the maximum and average difference between the measured and the predicted microhardness were respectively8.5%and3.2%.Applying the FNN model,the effects of the air pressure and processing duration on the microhardness at different depths were analyzed,revealing the significant interaction between the refined layer shelling and the continuous grain refinement.展开更多
Hardware neural networks controlled rotational actuators and application to an insect type micro robot are reported in this paper. Millimeter size rotational actuators are fabricated by combining MEMS （Micro Electro ...Hardware neural networks controlled rotational actuators and application to an insect type micro robot are reported in this paper. Millimeter size rotational actuators are fabricated by combining MEMS （Micro Electro Mechanical System） technology and shape memory alloy based artificial muscle wires. The actuator is composed of a pair of disk rotators and each rotor is suspended by four artificial muscle wires that are connected to the silicon frame. The rotational motion is generated by flowing the electrical current to each wire successively. Two actuators of different sizes are fabricated. The large actuator shows the displacement of 0.5 mm at the cycle time of 4 s. The small actuator shows 0.3 mm at 2 s. For controlling the actuator, the hardware neural networks are used. The hardware neural networks are composed of electrical circuits imitating cell bodies, excitatory synapses and inhibitory synapses. Four signal ports are extracted from four pairs of excitatory and inhibitory neurons and they are connected to the actuator. The small actuator is applied to the robot and built in the mid body of the robot. The shaft of the actuator is connected to the link mechanisms that transform the rotational motion to the locomotion. The appearance dimensions of the robot are 4.0, 2.7, 2.5 mm width, length and height. The robot performs forward and backward foot step like insects. The speed is 26.4 mm·min^-1 and the stepping width is 0.88 mm. Also, the robot changes the direction by external trigger pulses.展开更多
基金Supported by the Eu Information Technologies Programme Project(No. 22416) and National High Tech R&D Project(863/Computer Integrated Manufacture System AA413130) of China.
文摘A type of wavelet neural network, in which the scale function isadopted only, is proposed in this paper for non-linear dynamicprocess modelling. Its network size is decreased significantly andthe weight coefficients can be estimated by a linear algorithm. Thewavelet neural network holds some advantages superior to other typesof neural networks. First, its network structure is easy to specifybased on its theoretical analysis and intuition. Secondly, networktraining does not rely on stochastic gradient type techniques andavoids the problem of poor convergence or undesirable local minima.
基金Project (51475375) supported by the National Natural Science Foundation of China
文摘The bulk TC17was subjected to the high energy shot peening(HESP)at the air pressures ranging from0.35to0.55MPa and processing durations ranging from15to60min.The microhardness(HV0.02)from topmost surface to matrix of the HESP processed TC17was measured,which generally decreases with the increase of depth from topmost surface to matrix and presents different variation with air pressure and processing duration at different depths.A fuzzy neural network(FNN)model was established to predict the surface layer microhardness of the HESP processed TC17,where the maximum and average difference between the measured and the predicted microhardness were respectively8.5%and3.2%.Applying the FNN model,the effects of the air pressure and processing duration on the microhardness at different depths were analyzed,revealing the significant interaction between the refined layer shelling and the continuous grain refinement.
文摘Hardware neural networks controlled rotational actuators and application to an insect type micro robot are reported in this paper. Millimeter size rotational actuators are fabricated by combining MEMS （Micro Electro Mechanical System） technology and shape memory alloy based artificial muscle wires. The actuator is composed of a pair of disk rotators and each rotor is suspended by four artificial muscle wires that are connected to the silicon frame. The rotational motion is generated by flowing the electrical current to each wire successively. Two actuators of different sizes are fabricated. The large actuator shows the displacement of 0.5 mm at the cycle time of 4 s. The small actuator shows 0.3 mm at 2 s. For controlling the actuator, the hardware neural networks are used. The hardware neural networks are composed of electrical circuits imitating cell bodies, excitatory synapses and inhibitory synapses. Four signal ports are extracted from four pairs of excitatory and inhibitory neurons and they are connected to the actuator. The small actuator is applied to the robot and built in the mid body of the robot. The shaft of the actuator is connected to the link mechanisms that transform the rotational motion to the locomotion. The appearance dimensions of the robot are 4.0, 2.7, 2.5 mm width, length and height. The robot performs forward and backward foot step like insects. The speed is 26.4 mm·min^-1 and the stepping width is 0.88 mm. Also, the robot changes the direction by external trigger pulses.
