Human Workload Prediction: Lag Horizon Selection

TL;DR

This paper explores how different lag horizons affect the accuracy of human workload predictions in human-robot teams, revealing that univariate models need longer horizons than multivariate ones for optimal performance.

Contribution

It investigates the impact of variable lag horizons on univariate and multivariate time series models for human workload prediction, providing insights into optimal horizon lengths.

Findings

Univariate models require longer lag horizons (~240s) for accurate predictions.

Multivariate models perform well with shorter lag horizons (~120s).

Diminishing returns observed beyond 120s for multivariate models.

Abstract

Human-robot teams must be aware of human workload when operating in uncertain, dynamic environments. Prior work employed physiological response metrics from wearable sensors to estimate the current human workload; however, these estimates only enable robots to respond to under- or overload conditions reactively. Current human workload prediction approaches are limited to short prediction horizons and fail to investigate variable lag horizons' impact on predictions. This letter investigates the impact of lag horizons on both univariate and multivariate time series forecasting models for human workload prediction. A key finding is that univariate predictions required longer lag horizons of 240 seconds (s), whereas multivariate workload predictions sufficed with shorter lag horizons with diminishing returns around 120s.