Task phase recognition and task progress estimation for highly mobile workers in large building complexes

We evaluated the proposed methods in a large hospital complex, where the highly mobile workers were recruited among the non-clinical workforce. The evaluation is based on manually labeled real-world data collected over 4 days of regular work life. The collected data yields 83 tasks in total involving 8 different employees of a major university hospital with a building area of 160,000 m2. The results show that the proposed methods can with reasonable accuracy i) distinguish between the four most common task phases present in the workers' routines, achieving F1-Scores of 89.2%, and ii) estimate the task progress, yielding a mean error of 126.82 seconds for estimating the time until task completion and of 9.49 pp for estimating task progress.