Abstract: Addressing the issue that electric buses, due to their high motor torque and rapid braking response, can generate instantaneous accelerations during starting, stopping, and turning that easily exceed the human standing balance threshold, leading to passenger injuries inside the vehicle, this paper designs and implements a real-time ride comfort monitoring and active warning system for electric buses based on the ESP32 microcontroller. The system uses the high-performance ESP32- S3R8N8 microcontroller as the core processing unit, collects vehicle dynamic data through the ICM-42688 six-axis motion sensor, and combines a hybrid model of random forest (RF) and long short-term memory (LSTM) network to intelligently identify dangerous driving behaviors such as sudden braking, rapid acceleration, and sharp turns. It establishes a technology chain of "data collection–feature extraction–behavior recognition–voice warning". Vehicle tests show that the system can accurately capture high-risk driving behaviors and provide real-time alerts, forming an effective safety loop. This achievement provides a reliable technical solution for reducing passenger injuries in electric buses and lays a solid technical foundation for promoting the deep integration of electric buses with intelligent transportation and supporting the digital construction of smart bus systems.

Key words: electric bus; smoothness monitoring; ESP32-S3R8N8 microcontroller; ICM-42688 sixaxis motion sensor; random forest; long short-term memory network

摘要： 针对电动公交车因其电机扭矩大、制动响应快的特性，导致车辆在启停、转弯时产生的 瞬时加速度极易超出人体站立平衡阈值，致使乘客在车厢内摔伤这一突出问题，文章设计实 现了一套基于 ESP32 单片机的电动公交平顺度实时监测与主动预警系统。该系统以 ESP32-S3R8N8 高性能微控制器作为核心处理单元，通过 ICM-42688 六轴运动传感器采集车 辆动态数据，结合设计的随机森林（RF）与长短期记忆网络（LSTM）的混合模型，实现对 急刹车、急加速、急转弯等危险驾驶行为的智能识别，构建“数据采集－特征提取－行为识 别－语音预警”的技术链路。实车测试表明系统可精准捕捉高风险驾驶行为并实时告警，形 成有效的安全闭环。该成果为降低电动公交客伤事故提供了可靠技术方案，更为推动电动公 交车与智慧交通的深度融合、助力智慧公交体系的数字化建设奠定了坚实的技术基础。

关键词: 电动公交车；平顺度监测；ESP32-S3R8N8 单片机；ICM-42688 六轴运动传感器；随 机森林；长短期记忆网络