Purpose

In recent years, the demand for road traffic has continued to increase, but the casualties and economic losses caused by traffic accidents have also remained high. Therefore, the use of social service robots to manage, supervise and warn real-time traffic information has become an inevitable trend of traffic safety management.

Design/methodology/approach

In order to explore the inherent objective development law of road traffic accidents, in this paper, the factor analysis (FA) is used to explore the main influencing factors of traffic accidents, then the random forest algorithm is applied to build an FA–RF-based road traffic accident severity prediction model to predict two- and three-category accidents.

Findings

By comprehensively comparing the classification results of the two- and the three-category accident prediction, it also finds that due to the intersection between injuries and fatalities and the lack of necessarily external environmental information, the FA–RF model has a large degree of misjudgment for injuries and fatalities. Therefore, it is recommended to establish a real-time autonomous information communication mechanism between different kinds of social robots, which can improve the prediction of traffic accidents.

Originality/value

(1) A fusion model of FA–RF is considered to predict traffic accidents, which can be applied in traffic service robot. (2) It is recommended to establish a real-time autonomous information communication mechanism between different kinds of social robots, which can improve the prediction of traffic accidents.

中文翻译：

---

一种交通服务机器人道路交通事故预测模型

目的

近年来，道路交通需求不断增加，但交通事故造成的人员伤亡和经济损失也居高不下。因此，利用社会服务机器人对实时交通信息进行管理、监督和预警已成为交通安全管理的必然趋势。

设计/方法/方法

为了探索道路交通事故的内在客观发展规律，本文采用因子分析法（FA）探索交通事故的主要影响因素，然后应用随机森林算法构建FA-RF-基于道路交通事故严重程度预测模型预测二、三类事故。

发现

通过综合对比二类和三类事故预测的分类结果，还发现由于伤亡交叉，缺乏必要的外部环境信息，FA-RF模型存在较大程度的误判。受伤和死亡。因此，建议在不同种类的社交机器人之间建立实时自主的信息交流机制，可以提高对交通事故的预测能力。

原创性/价值

(1) 考虑采用FA-RF融合模型预测交通事故，可应用于交通服务机器人。(2) 建议在不同种类的社交机器人之间建立实时自主的信息交流机制，提高对交通事故的预测能力。