Energy Consumption Estimation in Electric Vehicles Considering Driving Style

Estimation of energy consumption of electric vehicles using Deep Convolutional Neural Network to reduce driver’s range anxiety

“…cov is the second best model after CNN 7 cov , in terms of RMSE, MAE and Corr. Values of RMSE, MAE and Corr can not be calculated for [18,19] as the techniques presented in these do not give real-time power/energy consumption as output and provide only single value of total energy consumption for the trip. It can be observed that all the approaches performed better on DS − II than on DS − I val .…”

“…The energy consumption was calculated by integrating the power over the time period. As [18,19] do not provide real-time power/energy consumption as output so, it was not possible to plot them. In Figure 12, each column represent the different road grade profile and each row represent the different drive cycle i.e., from top to bottom, rows represent UDDS, SFTP, HWFET and NEDC drive cycles, respectively and from left to right, columns represent Grade Profile 1 (constant grade at 0% i.e., no change in elevation), Grade Profile 2 (varies from -2% to 2%) and Grade Profile 3 (varies from -20% to 15%), as shown in Figure 11.…”

“…The main reason for that is the non-linear relationship of the influencing parameters which the MLR model was not able to estimate accurately as compared to the CNN 7 cov . The NN architecture presented in [18,19] have no hidden layer and has one input layer with 14 and 137 inputs, respectively and 1 output corresponding to the inputs. As the networks were shallow so they were also not able to accurately represent the non-linear relationship between the influencing factors.…”

“…For this, the neural networks were given input of vehicle speed, acceleration, and jerk. Felipe et al [19] extended [18] by adding route information to the input of the neural network. The ANN developed in [18,19], gave only one output of total energy consumed for the trip at the end of the trip.…”

Energy Planning for Unmanned Over-Actuated Road Vehicle

“…Many models have been developed for the power consumption estimation for electric or hybrid vehicles using neural network. In [8], a neural network based prediction of energy consumption is developed for an electric vehicle considering the driving style of driver. Sakayori et al [9] used multilayer perception (MLP) neural network to predict power consumption for an autonomous mobile robot, including the slope, heading angle, and velocity profile as inputs.…”