Enhanced vision-transformer integrating with semi-supervised transfer learning for state of health and remaining useful life estimation of lithium-ion batteries

•An enhanced ViT with SSL-transfer learning is proposed for SOH & RUL estimation.•Depth separable convolution is introduced to capture local battery aging details.•Maximum mean discrepancy is used to reduce the inconsistency of data distribution.•SSL is designed with consistency regularization to label the abundant aging data.•SOH & RUL average errors were within 0.6 % and 3.86 % under 40 °C dynamic cycles.