In-field piecewise regression based prognosis of the IPC in electrically powered agricultural machinery

The energy consumption in electrically powered machinery (EPM) depends on the manoeuvres, the mass of the vehicle, its load, the characteristics of the terrain, the deformation of the wheel, the slippage, the ambient and batteries temperature, among other issues, changing the instantaneous power consumption (IPC) behaviour. An accurate estimate of energy consumption (and therefore, of the IPC) will lead to an efficient battery recharging strategy. To overcome the IPC unmodelled issues previously mentioned, this work presents a procedure for predicting the energy consumed by EPMs through IPC prognosis, tested and validated on three different terrain types: gravel, clay and pavement. To this end, a fixed polynomial model of the IPC with respect to the terrain type is obtained as a priori knowledge. Then, through new readings of the IPC, the model is updated by segments and later used for IPC prognosis given a previously defined route. The experimental results show an improvement in the estimation of energy consumption (and therefore, of the energy still available for traversing) of 56.22% with respect to the data provided by the manufacturer and of 7.14% compared to theoretical and empirical approaches previously published. Although tested in agricultural scenarios, the methodology presented here encourages to be applied in other contexts of electro-mobility since it offers a suitable technique for better managing operational costs.