| Multi-signal Bigdata |
| Multi-signals that can be received by smartphones are converted into big data in real time, and signals that can be received indoors can be used for positioning in dozens of types as shown in the figure. |
| Generating Sensor Map |
| Smartphones receive many sensor information like as stars in the night sky, map them to the location information, and divide them into sensors/coordinates/indexes to create a sensor map for deep learning that can be learned in real time. |
| Deep Learning PNN (Horizontal positioning) 1.74m Horizontal Accuracy |
| Sensor maps for deep learning exist in similar forms throughout the room Real-time positioning of similar sensor maps by dividing the similarity between PNN deep learning sensor maps to which the constellation algorithm is applied. It was impossible to implement with general deep learning technology CNN |
| Deep Learning PNN (vertical positioning) 99.2% verticalAccuracy |
| "Relative pressure map" base precision vertical positioning Development of "EM(Expectation-maximization)algorithm" based optimization method to compensate for changes in atmospheric pressure in real time. Real-time update of actual barometric pressure Guarantee sustainable vertical accuracy |
