An Improved ECM-Based State-of-Charge Estimation for SLA and LFP Batteries Used in Low-Cost Agricultural Mobile Robots

Abstract

Batteries are crucial in transitioning from fossil fuels to clean-powered mobility, for several applications such as Electric Vehicles and Agricultural Mobile Robots (AMRs). However, the adoption of AMRs is limited by several challenges related to battery management, including restricted operation time, long recharge periods, and safe operation. The State of Charge (SOC) provides information about the remaining energy in the battery and is essential for battery management. Therefore, an accurate SOC estimation is crucial to ensure safe and reliable operation, which is needed to overcome the aforementioned challenges. This paper proposes, implements, and validates an SOC estimation system for low-cost AMRs. The accuracy of the SOC estimation is improved by adding information about the battery’s Open Circuit Voltage (OCV) to the Equivalent Circuit Models (ECM). Two SOC estimation methods based on ECM were implemented and validated for a Lithium Iron Phosphate battery (LFP) and a Sealed Lead Acid (SLA) battery powering an AMR. Finally, the results indicate that adding the OCV information to the models improves the estimation accuracy for both chemistries, being particularly interesting for LFP batteries, whose OCV vs. SOC has a flat area in almost the entire useful region of the SOC.

Keywords

• Agricultural robot
• state of charge
• battery
• Kalman filter
• parameter estimation