Neural modeling and brain-based devices using special purpose processor
A special purpose processor (SPP) can use a Field Programmable Gate Array (FPGA) or similar programmable device to model a large number of neural elements. The FPGAs can have multiple cores doing presynaptic, postsynaptic, and plasticity calculations in parallel. Each core can implement multiple neural elements of the neural model.
1. A robot device including:
a) multiple removable pods including at least one sensor, and a wheel controlled by a motor, the multiple removable pods including a bi-directional suspension; and
b) a central axis operably connected to the bi-directional suspension of the multiple removable pods.
2. The robot device of claim 1 , further comprising;
a computing element implementing a neural model, the neural model including multiple neural elements, the computing element including multiple cores to implement neural elements, the cores processing data in parallel, the processing in the cores including presynaptic calculations using input values and weights, postsynaptic calculations to produce postsynaptic outputs using the results of the presynaptic calculations and plasticity calculations to modify the weights.
3. A robot device including:
a) multiple pods including at least one sensor and at least one actuator, at least two of the multiple pods including a wheel controlled by a motor, wherein the pods are removable and have a bidirectional suspension;
b) a central portion including at least one motorized tread; and
c) a computer implemented neural model including multiple neural elements, the computer implemented neural model including cores to implement the neural elements, the cores processing data in parallel, the processing in the cores including presynaptic calculations using input values and weights, postsynaptic calculations to produce postsynaptic outputs using the results of the presynaptic calculations and plasticity calculations to modify the weights.