Self-learning intelligent driving device
A self-learning intelligent driving device including: a first neural network module for performing a corresponding action evaluation operation on an input image to generate at least one set of trajectory coordinates: a switching unit controlled by a switching signal, where when the switching signal is active, data received at a first port is sent to a second port, and when the switching signal is inactive, data received at the first port is sent to a third port; a second neural network module for performing a corresponding image evaluation operation on the at least one set of trajectory coordinates when the switching signal is active to generate at least one simulated trajectory image; and a driving unit having a robotic arm for generating at least one corresponding motion trajectory according to the at least one set of trajectory coordinates when the switching signal is inactive.
1. A self-learning intelligent driving device, including:
a first neural network module for performing a corresponding action evaluation operation on an input image to generate at least one set of trajectory coordinates;
a switching unit; having a first port, a second port, and a third port, and being controlled by a switching signal, where the first port is used to receive the at least one set of trajectory coordinates, when the switching signal is in an active state, data received at the first port will be sent to the second port, axed when the switching signal is in an inactive state, data received at the first port will be sent to the third port;
a second neural network module for performing a corresponding image evaluation operation on the at least one set of trajectory coordinates to generate at least one simulated trajectory image when the switching signal is in the active state; and
a driving unit having a robotic arm for generating at least one corresponding motion trajectory according to the at least one set of trajectory coordinates when the switching signal is in the inactive state.
2. The self-learning intelligent driving device as disclosed in claim 1 , further including a user operation interface, wherein the user operation interface has a learning process activation icon for a user to press, and when the learning process activation icon is pressed, the user operation interface will cause the switching signal to be in the active state.
3. The self-learning intelligent driving device as disclosed in claim 1 ; wherein a plurality of weighting coefficients of the second neural network module are derived by using plural data sets of (the trajectory coordinates, an image of the motion trajectory) to perform a first deep learning training process.
4. The self-learning intelligent driving device as disclosed in claim 3 , wherein a plurality of weighting coefficients in the first neural network module are updated by a method, which includes: connecting the first neural network module with the second neural network module and using plural data sets of (the input image, the simulated trajectory image) to perform a second deep learning training process, and only adjusting the weighting coefficients of the first neural network module but not adjusting the weighting coefficients of the second neural network module during the second deep learning training process.
5. The self-learning intelligent driving device as disclosed in claim 1 , wherein the input image is derived from a calligraphy copybook.