Long Term Motion Prediction Using Keyposes

Long term human motion prediction is essential in safety-critical applications such as human-robot interaction and autonomous driving. In this paper, we show that, to achieve long term forecasting, predicting human pose at every time instant is unnecessary. Instead, it is more effective to predict a few keyposes and approximate intermediate ones by linearly interpolating the keyposes. We will demonstrate that our approach enables us to predict realistic motions for up to 5 seconds in the future, which is far larger than the typical 1 second encountered in the literature. Over this extended time period, our predictions are more realistic and better preserve the motion dynamics than those state-of-the-art methods yield. Furthermore, because we model future keyposes probabilistically, we can generate multiple plausible future motions by sampling at inference time. This is useful to model because people usually can do one of several things given what they have already done.