Research Article
Online Traffic Accident Spatial-Temporal Post-Impact Prediction Model on Highways Based on Spiking Neural Networks
| Attribute category | Attributes | Value and description |
| Accident attribute | Id | Unique identifier of the accident record | TMC event code | 202 serious accidents | 203 multivehicle accidents | 245 two-lane blocked | 246 three-lane blocked | …… | Block lanes | The number of lanes affected by the accident | Block distance | The length of the road extent blocked by the accident |
| Spatial attribute | Latitude | Latitude in GPS coordinate of the accident point | Longitude | Longitude in GPS coordinate of the accident point | County | County in address field | Motorway ID | Motorway name in address field | Sensor ID | Unique identifier of the nearest sensor to the accident point |
| Temporal attribute | Day or night | 0 day | 1 night | Workday or not | 0 workday | 1 holiday or weekend | Peak hour or not | 0 morning peak hour | 1 evening peak hour | 2 normal period |
| Weather attribute | | 0 sunny | 1 rain | 2 snow | 3 thunderstorm | 4 fog |
| Traffic flow characteristic attribute | Flow data series | A matrix of flowrate of upstream sensors after accident happens | Speed data series | A matrix of speed of upstream sensors after accident happens |
| Impact attribute | Accident clean-up time | Sum of detection, verification, response, and clearance time | Full-recovery time | Time duration from clearance to the moment when no vehicle is accumulated | Half-recovery time | Time duration from clearance to the moment when queue length is half of maximum accumulative queue length | Maximum accumulative queue length | — | Average accumulative queue length | — |
|
|