CovarianceNet: Conditional Generative Model for Correct Covariance Prediction in Human Motion Prediction

Postnikov, Aleksey; Gamayunov, Aleksander; Ferrer, Gonzalo

doi:10.1109/iros51168.2021.9635968

Cited by 4 publications

(1 citation statement)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2) Final Displacement Error (FDE): 2 distance between the predicted mean final position and the GT final position at the prediction horizon T . 3) Delta Empirical Sigma Value (∆ESV i ) [55]: The difference in the fraction of GT positions that fall within the i-σ level set (e.g., 1σ, 2σ, 3σ) of the predicted distribution and the fraction from an ideal Gaussian. In particular, ∆ESV i is a useful metric for identifying overor underconfidence, as ∆ESV i := σ pred,i − σ ideal,i where σ pred,i is the empirical fraction of GT positions that lie within the i-sigma level set of the prediction distribution and σ ideal,i is the expected fraction from a perfectly-calibrated bivariate Gaussian, where σ ideal,1 ≈ 0.39, σ ideal,2 ≈ 0.86, and σ ideal,3 ≈ 0.99 [56].…”

Section: Propagating State Uncertainty Through Trajectory Forecastingmentioning

confidence: 99%

Propagating State Uncertainty Through Trajectory Forecasting

Ivanovic¹,

Lin²,

Shrivastava³

et al. 2021

Preprint

View full text Add to dashboard Cite

Uncertainty pervades through the modern robotic autonomy stack, with nearly every component (e.g., sensors, detection, classification, tracking, behavior prediction) producing continuous or discrete probabilistic distributions. Trajectory forecasting, in particular, is surrounded by uncertainty as its inputs are produced by (noisy) upstream perception and its outputs are predictions that are often probabilistic for use in downstream planning. However, most trajectory forecasting methods do not account for upstream uncertainty, instead taking only the most-likely values. As a result, perceptual uncertainties are not propagated through forecasting and predictions are frequently overconfident. To address this, we present a novel method for incorporating perceptual state uncertainty in trajectory forecasting, a key component of which is a new statistical distance-based loss function which encourages predicting uncertainties that better match upstream perception. We evaluate our approach both in illustrative simulations and on large-scale, real-world data, demonstrating its efficacy in propagating perceptual state uncertainty through prediction and producing more calibrated predictions.

show abstract

Section: Propagating State Uncertainty Through Trajectory Forecastingmentioning

confidence: 99%