enhanced locomotion efficiency of a bio-inspired walking robot using contact surfaces with frictional anisotropy
Based on the principle of morphological calculation, we propose a new method, which uses the interaction between passive heterosexual scales. like material (e. g. , shark skin)and a non- Smooth substrate to improve the movement efficiency of the robot walking on the slope. Real robot experiments show that passive friction The enhanced surface of the robot\'s abdomen or feet enables the robot to grasp a specific surface and move effectively with reduced energy consumption. We have supplemented robot experiments to study the friction properties of shark skin and its mechanical stability. Due to a series of tilted teeth, it shows a high degree of friction to the opposite sex. The direction of the teeth towards the underlying collagen also strongly affects their mechanical binding to the substrate. This study not only opens up new ways to realize energy Efficient leg-type robot movement can also better understand the functional and mechanical properties of the opposite sex surface. This understanding will help to develop new materials for other entities World applications. All the methods and experiments reported here are carried out in accordance with the approved guidelines. In our experiment, the skin of a dead pig () The animal collection at the Institute of Animal Studies at the University of Kiel in Germany is used. Cut out a piece of skin from the shark\'s belly and keep it deep- Freeze at-20 °c 4 samples were sawed from frozen skin 20 × 40mm. Remove the fleshy part of the skin. Two samples were nailed to a piece of wood and dried under ambient conditions. Two other samples were stored in 2% diethylene glycol before the experiment. Replica of a sled made of wood made of polished paper with different roughness (0. 3, 1, 3, 12u2009m) Spurr epoxy is made of glue stuck to the bottom of the sled. In order to avoid the possible edge effects during the friction measurement, all edges of the sled are rounded in. The base area of the sled is 15 × 15mm. Negative replicas were made using dental wax ( President of ulsterten cortyne/Whaledent AG light body, Switzerland). An example of friction measurement and a scheme of experimental devices are given in and B, respectively, to determine the friction coefficient of shark skin in different directions under different roughness. A shark\'s skin sample is glued to a glass slide. Polish the positive copy of the paper with 0. Particle sizes of 3 m, 1 m, 3 m and 12 m ( Buehler, Lake Bluff, IL, United States of America) Used as a rough counterpart in a friction measurement. Spurr replicas of different roughness were placed on shark skin. Set the weight of 50g and 200g on the sled. S & P sledge is connected to the force sensor ( FORT1000, World precision instruments, Sarasota, FL, United States of America)using a non- Stretch nylon rope. The force sensor is mounted on a motorized manipulator that pulls a sled 10mm with 2. 5mm/s speed. The signal of the force sensor is collected by the Biopac system ( Biopac Systems Inc. Goleta, California, USA) And record using AcqKnowledge 3. 7. 0. software ( Biopac Systems Inc. Goleta, California, USA). Measure friction in three different directions: Tail ( Along the small teeth), rostral ( For fine teeth), and lateral; At different roughness (0. 3 m, 1 m, 3 m and 12 m); Under two different loads (50u2009g and 200u2009g); In the sea ( Water from the Baltic Sea, fresh unfrozen shark skin) In a dry state. For each parameter combination, friction measurements are performed at least three times. After three measurements in each group, check the skin for damage and take a new location for further measurements if needed. Coefficient of friction u2009 = u2009/is calculated using the highest pull force within the first 0. 5mm pull distance. Initial area of the force- The distance curve corresponds to the straightening of the thread connecting the base plate and the force sensor and the slight rotation of the base plate on the shark skin. However, we can verify that the start time of the sliding motion is within 0. 5u2009mm. All measurements are carried out under normal circumstances ( Room temperature, humidity 40%). Dry samples are coated with ~ 10 nm sputtering of Au/Pd. Imaging with Hitachi TM3000 desktop scanning electron microscope (Hitachi High- Technology company, Tokyo, Japan) Under the acceleration voltage of 3kv. Since the observed friction heterogeneity is not related to the substrate roughness, the data are collected accordingly. To test the significant differences in different sliding directions in dry and wet states, Way ANOVA has been carried out. Although the assumption of normal distribution and the same variance was violated, an analysis of variance was performed. Since the sample size of each group is roughly the same, the ratio of the maximum group variance to the maximum group variance is less than 10, so the ANOVA is considered robust enough. Also, Holm- After Sidak multiple comparisons After the event, a special test was conducted without assuming the same party difference. Statistical analysis was performed using SigmaPlot 12. 5 ( San Jose etiat Software Co. , Ltd. , California, USA). Neural control generated by biological movement AMOS was developed in early work. Here we used it in robotic experiments without any modifications and sensory feedback. The control consists of three neural modules: the central mode generator (CPG)- Control module based on neural modulation, neural CPG post-processing module and neural motion control module. The CPG- Different periodic signals are generated based on the control module to obtain different gaits. The post-processing module shapes the CPG cycle signal to obtain smooth leg movement. The motor control module consists of two other different networks. Phase switching network (PSN) And speed regulation network (VRNs)] Used to control the direction of walking ( Forward/backward and turn). The final output of the motor control module is transmitted through the extension cord to all leg joints of the AMOS. All neurons of the motion control network are modeled as discrete-time non-Pulse neurons They are Updated at about 27 hz. The activity of each neuron is developed according to the following: where it represents the number of units, the internal bias term, or a fixed input to the neuron. Synaptic strength of connections from neurons to neurons. The output of all neurons of the network is by using a two-surface tangent () Transfer function, I. e. , u2009=u2009tanu2009h(), [−1, 1] In addition to CPG postprocessing neurons using step functions, motor neurons using segmented linear transfer functions, and neurons using linear transfer functions for search and lifting control. A complete description of the motion control network can be seen in our previous work. The six- Legged is a bio-inspired hardware platform. It consists of six identical legs, each with three joints ( Three degrees of freedom): the thoraco-coxal (TC-) Joint enable forward (+)and backward (−) Sports, coxo-trochanteral (CTr-) Joint elevation (+) And depression (−) Legs, women\'s. tibial (FTi-) Connection enables extension (+)and flexion (−)of the tibia. This multi-form The connected legs mimic the cockroach legs, but the tar Bone part is ignored. All joints are driven by standard servo motors. All 19 motors and 32 sensors have walking machines. We use a Multi-Servo IO-Board (MBoard) Digitize all sensory input signals and generate pulseswidth- Control the modulation signal of the servo motor position. For the robot walking experiment in this study, the MBoard is connected to a personal computer on which the neural motion controller is implemented. The update frequency is 27 hz. The battery provides power supply: 1 11. All servo motors and 1v lithium polymer 3200mah for two 11. 1 Electronic Board of lithium polymer 910 mAh (MBoard) All sensors (see ref. More details).