Prediction of Remaining Useful Life of the Space Rover Battery Using Fuzzy Adaptive Neural Learning Network

Space rovers operate in extreme and unpredictable environments, requiring advanced systems for real-time health monitoring and failure prediction. This research presents a hybrid approach that integrates Fuzzy Logic and Artificial Neural Networks (ANNs) to enhance the reliability and safety of rover operations, with a particular focus on predicting the Remaining Useful Life (RUL) of the rover’s battery. The Fuzzy Logic system deals with the uncertainties in the sensor data due to the rough conditions on Mars and assesses the environmental conditions, for instance, wind speed, temperature, and battery status while defining fuzzy sets and rules based on these variables for the overall risk-level assessment. At the same time, an ANN will be used to process the sequence of sensor data and predict the RUL of the battery. The ANN, trained on historical data, learns patterns that correlate environmental factors and system stresses with potential battery failures, providing predictive insights for proactive maintenance. This model meets its stated accuracy level and has performance metrics that indicate that the predicted RUL values are highly correlated to the actual RUL values. The Fuzzy Logic system provides the necessary comprehension for the risk assessment during a mission, while ANN can provide accurate predictions for the future ensuring that reliability is maintained. Adoption of these two approaches, FANLN for example, will enhance the rover’s capability to adapt to changing Martian environmental conditions, reduce risks effectively, and extend mission durations improving Martian exploration and colonization strategies.