IMU 기반 3륜 전동차의 사이드슬립각 최적화 연구 = Optimization of Sideslip Angle Estimation for a Three-Wheeled Electric Vehicle Using an IMU

다국어 초록 (Multilingual Abstract)

This study analyzes the limitations of IMU (Inertial Measurement Unit)-based velocity estimation for sideslip angle estimation in a three-wheeled electric vehicle. Due to the asymmetric structure and narrow rear track, three-wheeled vehicles are highly susceptible to lateral instability during cornering, and accurate estimation of longitudinal and lateral velocity is a prerequisite for reliable sideslip angle estimation. To evaluate the applicability and inherent limitations of using a single low-cost IMU, an acceleration-integration-based velocity estimation algorithm was implemented on an embedded vehicle control unit (VCU) and experimentally validated using an RT3002 GNSS/INS reference system.
A front-steering, rear-wheel-drive three-wheeled electric vehicle was used as the test platform. The IMU signal processing pipeline consisted of offset correction, zero-velocity update (ZUPT), low-pass filtering, and trapezoidal integration. Straight-line driving tests showed that the IMU–VCU-based velocity estimate followed the RT3002 reference trend, indicating that basic velocity estimation is feasible under mild conditions.
However, S-shaped curve-driving tests revealed structural limitations of IMU-only velocity estimation. The lateral velocity exhibited asymmetric drift during left and right turns, and the longitudinal velocity displayed bias accumulation and discontinuous variations. These errors were primarily caused by IMU–body-frame misalignment, gravity projection due to sensor tilt, and the absence of yaw-rate compensation for the rotational acceleration term (r·Vx). Consequently, pure acceleration integration approaches inherently suffer from cumulative bias and asymmetric errors during cornering maneuvers.
The results explicitly quantify the practical limitations of single-IMU-based velocity estimation for sideslip angle estimation in three-wheeled electric vehicles. The findings provide a technical basis for future work involving yaw-rate and gravity compensation, coordinate alignment, and sensor-fusion-based state estimation algorithms.

목차 (Table of Contents)

• Ⅰ. 서론 1
• Ⅱ. 연구방법 및 실험구성 5
• 2.1 하드웨어 구성 5
• 2.2 사이드슬립각 추정 알고리즘 설계 17
• 2.3 실험 시나리오 및 검증 절차 21
• Ⅲ. 실험결과 및 분석 23
• 3.1 실험 환경 구성 23
• 3.2 실험 환경 구축 및 수행 절차 25
• 3.3 속도 데이터 분석 및 비대칭 현상 고찰 30
• 3.4 이상현상 원인 분석 32
• Ⅳ.결론 36
• 참고문헌 37
• ABSTACT 38