Predictive Thin Plate Spline Model for Estimation of Load Carriage at Varying Gradient and Speed

Abstract

Laboratory-based experimental studies are being carried out for Indian soldiers to estimate optimal load carriage at different gradients and walking speeds. These experiments involve the recording of Cardio-respiratory responses such as Heart Rate (HR), Oxygen Consumption (VO2), Energy Expenditure (EE), Respiratory Frequency (RF), Minute Ventilation (VE), and Maximal Aerobic capacity (%VO2max). Due to limitations in the data sample size that can be obtained in laboratory-based experiments, there is a need for mathematical interpolation to obtain intermediate values in the study. Load carriage can be affected by factors that can be controlled, such as the speed of marching, and also by external factors that cannot be controlled like ambient temperature. Real-time interactions of all the factors also have an impact on the load-carrying capacity. Planning of the mission operations requires the specification of well-defined work-rest schedules and indication of total load limits, to ensure the operational effectiveness of the military personnel. In this paper, we present a Predictive 3-Dimensional Thin Plate Spline Model for efficient estimation of load. We developed a Multiple Linear Regression Model for predicting %VO2max for combinations of load and gradient. The accuracy of the model was 85 per cent and the maximum permissible loads were derived from the prediction model for the physiological limits of 50 per cent, 60 per cent, and 75 per cent of VO2max. A Thin Plate Spline based interpolation technique was used on this Multiple Linear Regression Model to generate optimal load at intermediate values for the experimental study. A similar predictive Interpolation Model was also developed for estimating load for varying walking speeds at level ground.