Trajectory prediction method, device, electronic equipment and readable storage medium

Abstract

The embodiment of the invention provides a trajectory prediction method, a device, electronic equipment and a readable storage medium, and relates to the technical field of computers. Through the embodiment of the invention, a long-short-term memory model can sequentially predict prediction coordinates corresponding to each first time node according to a time sequence based on a plurality of neuronal units connected in series in sequence, and in the process, the prediction coordinates of each first time node can be obtained. Since the long-term and short-term memory model can well learn the correlation between the time series data and alleviate the problem of time dependence, the prediction coordinates of each first time node in the embodiment of the invention are determined by the neuron unit based on the prediction coordinates output by the previous neuron unit, i.e., the prediction coordinates of each second time node are determined by the neuron unit based on the prediction coordinates output by the previous neuron unit. According to the embodiment of the invention, the prediction coordinates of a plurality of subsequent first time nodes can be predicted on the basis of the initial target characteristic state, so that the prediction coordinates have relatively strong correlation, and furthermore, the long-term and short-term memory model can still output an accurate prediction track as time goes on.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a trajectory prediction method, device, electronic equipment and readable storage medium. Background technique [0002] At present, in order to provide better services, service platforms often conduct simulations in advance through simulation systems. For example, platform A is a first-generation driving service platform, and platform A needs to allocate Orders, in order to make the distribution of orders more reasonable, it is necessary to simulate the distribution of orders through the simulation system. [0003] In the process of simulation, in order to make the simulation system accurately reflect the actual situation, the simulation system needs to predict the walking path of each substitute driver (especially the idle driver). In related technologies, a random probability matrix or However, as time goes on, the accuracy of the above method will become lower and...

AI Technical Summary

Problems solved by technology

[0003] In the process of simulation, in order to make the simulation system accurately reflect the actual situation, the simulation system needs to predict the walking path of each substitute driver (especially the idle driver). In related technologies, a random probability matrix or However, as time goes on, the accuracy of the above method will become lower and lower, which will lead to the worsening of the simulation effect of the simulation system.

Images