Trying to reach for a nearby object is a mindless task for an average person, but the same task requires a complex neural network that costs countless years to evolve. In recent times, robots also acquired the same ability by leveraging artificial neural networks. According to a recent study, robot hands need to learn almost three different motions to grasp objects off a surface. The main aspect which fueled this innovation is called spiking neuron, just like how in human brain real neuron work, artificial neurons are used by robots. Researchers are now closely studying SNN to understand the approach to yield positive insights on how neural works.
A research scientist at FZI Germany, Juan Camilo Vasquez Tieck, stated that the programming that is needed for bio-inspired robot or humanoid is very sophisticated. He furthermore added that the conventional methods to develop new capabilities are not always a viable option. Traditional robotic systems make use of thousands of extensive calculations to detect trajectories and grab the object. But the new robotic system like Tieck’s will function using SNN. In this new method, the neural will be trained first, and then the model system will be polished along with object motions.