BACKGROUND Policymakers need accurate data to develop efficient interventions to promote transport physical activity. Given the imprecise assessment of physical activity in trips, our aim was to illustrate novel advances in the measurement of walking in trips, including in trips incorporating non-walking modes. METHODS We used data of 285 participants (RECORD MultiSensor Study, 2013-2015, Paris region) who carried GPS receivers and accelerometers over 7 days and underwent a phone-administered web mobility survey on the basis of algorithm-processed GPS data. With this mobility survey, we decomposed trips into unimodal trip stages with their start/end times, validated information on travel modes, and manually complemented and cleaned GPS tracks. This strategy enabled to quantify walking in trips with different modes with two alternative metrics: distance walked and accelerometry-derived number of steps taken. RESULTS Compared with GPS-based mobility survey data, algorithm-only processed GPS data indicated that the median distance covered by participants per day was 25.3 km (rather than 23.4 km); correctly identified transport time vs. time at visited places in 72.7% of time; and correctly identified the transport mode in 67% of time (and only in 55% of time for public transport). The 285 participants provided data for 8983 trips (21,163 segments of observation). Participants spent a median of 7.0% of their total time in trips. The median distance walked per trip was 0.40 km for entirely walked trips and 0.85 km for public transport trips (the median number of accelerometer steps were 425 and 1352 in the corresponding trips). Overall, 33.8% of the total distance walked in trips and 37.3% of the accelerometer steps in trips were accumulated during public transport trips. Residents of the far suburbs cumulated a 1.7 times lower distance walked per day and a 1.6 times lower number of steps during trips per 8 h of wear time than residents of the Paris core city. CONCLUSIONS Our approach complementing GPS and accelerometer tracking with a GPS-based mobility survey substantially improved transport mode detection. Our findings suggest that promoting public transport use should be one of the cornerstones of policies to promote physical activity.

中文翻译：

---

将传感器跟踪与基于GPS的移动性调查相结合，可以更好地测量出行中的身体活动：公共交通产生步行。

背景技术决策者需要准确的数据来开发有效的干预措施以促进运输身体活动。考虑到对旅行中身体活动的不精确评估，我们的目的是说明在旅行中步行测量中的新进展，包括在采用非步行模式的旅行中。方法我们使用了285名参与者（RECORD MultiSensor研究，2013-2015，巴黎地区）的数据，这些参与者携带了7天的GPS接收器和加速度计，并在经过算法处理的GPS数据的基础上进行了电话管理的网络移动性调查。通过此流动性调查，我们将旅行分为开始/结束时间，经过验证的旅行模式信息以及手动补充和清洁的GPS轨迹，将其分解为单峰旅行阶段。通过此策略，您可以通过两种替代指标来量化不同模式下的步行：步行距离和加速度计得出的步数。结果与基于GPS的流动性调查数据相比，仅通过算法处理的GPS数据表明，参与者每天的平均距离为25.3 km（而不是23.4 km）；在72.7％的时间内正确识别了运输时间与参观地点的时间; 并在67％的时间内正确识别了运输方式（仅在55％的时间内用于公共运输）。285名参与者提供了8983次旅行的数据（21,163个观察段）。参与者在旅行中花费的时间中位数为总时间的7.0％。每次步行的平均步行距离为0.40 km，公共交通的步行平均距离为0.85 km（相应行程的加速度计步长的中位数分别为425和1352）。总体而言，旅行的总距离中有33.8％是步行旅行，而37。在公共交通旅行中，累计了3％的加速度计步数。与巴黎核心城市的居民相比，在每8小时的穿戴时间中，远郊居民的每日行走距离累计低1.7倍，而旅途中的步数减少了1.6倍。结论我们的方法通过基于GPS的移动性调查补充了GPS和加速度计跟踪，从而大大改善了运输模式检测。我们的发现表明，促进公共交通的使用应成为促进体育锻炼的政策的基石之一。结论我们的方法通过基于GPS的移动性调查补充了GPS和加速度计跟踪，从而大大改善了运输模式检测。我们的发现表明，促进公共交通的使用应成为促进体育锻炼的政策的基石之一。结论我们的方法通过基于GPS的移动性调查补充了GPS和加速度计跟踪，从而大大改善了运输模式检测。我们的发现表明，促进公共交通的使用应成为促进体育锻炼的政策的基石之一。