code and artificial neural network is proposed and verified. A number of LEDs are placed above

the detection area，and a number of photoelectric detection units are randomly arranged below.

The distribution of light intensity caused by blocking light can be regarded as the characteristic

of spatial position. The light intensity distribution data is processed and learned

by using the pre-built neural network to realize the mapping relationship between the light

intensity distribution and the spatial coordinates of the object，so as to realize the spatial

positioning of the object with high precision and fast response. In the experiment，6 LEDs and 8

photovoltaic cells are used to build a demonstration system，and spatial positioning is carried out

through actual measurement.