Game controller with touchpad input
A game controller includes a touchpad that a user, viewing a virtual keyboard on a screen, can soft-touch to move a cursor on the screen and then hard-touch to move the cursor and also send location data to a processor for inputting a letter from the virtual keyboard. Machine learning is used to predict a next letter or next word.
1. An apparatus, comprising:
at least one computer storage that is not a transitory signal and that comprises instructions executable by at least one processor to:
receive a touch signal on a touch pad of a computer simulation controller;
responsive to the touch signal indicating a first pressure, move a cursor on a display distanced from the controller and not establish a selection of a letter;
responsive to the touch signal indicating a second pressure greater than the first pressure, establish a selection of a first alpha-numeric character and present the first alpha-numeric character on the display;
input the first alpha-numeric character to at least a first neural network (NN); and
receive from the first NN a predicted sequence of alpha-numeric characters comprising at least a first predicted alpha-numeric character, wherein the first NN comprises plural long short-term memory (LSTM) networks.
2. The apparatus of claim 1 , wherein the processor is embodied in the computer simulation controller.
3. The apparatus of claim 1 , wherein the processor is embodied in a computer simulation console configured for communicating with the computer simulation controller.
4. The apparatus of claim 1 , wherein the instructions are executable to:
responsive to the touch signal indicating the second pressure, move the cursor on the display.
5. The apparatus of claim 1 , wherein the instructions are executable to:
responsive to the touch signal indicating the first pressure, enlarge an image of a keyboard on the display.
6. The apparatus of claim 1 , wherein the instructions are executable to:
present on the display, next to the first alpha-numeric character, the predicted sequence of alpha-numeric characters comprising at least the first predicted alpha-numeric character.
7. A method comprising:
receiving, from a computer simulation controller, a touch signal;
responsive to the touch signal indicating a first pressure, moving a cursor on a display;
responsive to the touch signal indicating a second pressure greater than the first pressure, establishing at least a first letter at least in part using heat map statistics, the heat map statistics representing a path of a finger swipe with probabilities of each letter associated with the path being identified at time intervals along the swipe, wherein at each time interval, a letter with a highest probability is identified as the first letter; and
presenting the first letter on the display.
8. The method of claim 7 , comprising:
inputting the first letter to at least a first neural network (NN); and
presenting on the display at least a first predicted letter generated by the first NN.
9. The method of claim 8 , comprising:
presenting on the display at least a sequence of predicted letters comprising the first predicted letter generated by the first NN.
10. The method of claim 8 , comprising:
inputting the first NN words previously input to the computer simulation controller usable by the first NN to generate the first predicted letter.
11. A method comprising:
receiving, from a computer simulation controller, a touch signal;
responsive to the touch signal indicating a first pressure, moving a cursor on a display;
responsive to the touch signal indicating a second pressure greater than the first pressure, establishing at least a first letter;
inputting the first letter to at least a first neural network (NN);
presenting on the display at least a first predicted letter generated by the first NN; and
training the first NN using text input by computer simulation players on a network.
12. The method of claim 11 , wherein the training comprises processing the text input by computer simulation players using at least one bidirectional long short-term memory (LSTM) network.
13. A method comprising:
receiving, from a computer simulation controller, a touch signal;
responsive to the touch signal indicating a first pressure, moving a cursor on a display;
responsive to the touch signal indicating a second pressure greater than the first pressure, establishing at least a first letter;
inputting the first letter to at least a first neural network (NN);
presenting on the display at least a first predicted letter generated by the first NN; and
inputting to the first NN words from a simulation dictionary usable by the first NN to generate the first predicted letter.
14. A method comprising:
receiving, from a computer simulation controller, a touch signal;
responsive to the touch signal indicating a first pressure, moving a cursor on a display;
responsive to the touch signal indicating a second pressure greater than the first pressure, establishing at least a first letter;
inputting the first letter to at least a first neural network (NN);
presenting on the display at least a first predicted letter generated by the first NN; and
inputting to the first NN data from a character sequence model generated by at least one bidirectional long short-term memory (LSTM) network and usable by the first NN to generate the first predicted letter.