The Food Friend project consists of a consortium of 5 countries and 17 partners. The smart health project will run from October 2019 to October 2023. The goal of the Food Friend project is to develop a complete toolset, consisting of hardware, software, and methodologies, that can automatically measure a person’s food intake with a minimum of required user input and turn it into personalized and actionable feedback. The technologies are developed for use by care professionals, research institutions, caterers, or home users to get a better overview of a person’s dietary behavior.

Development goals

This project merges several initiatives around software development and metabolic data processing in the following initiatives, which we will discuss in upcoming articles.

  • MOX Accelerometry data processing software;
    • Batch processing accelerometer sensor data
    • Integrating new outcome measures for sedentary behavior, sleep / wake time, bouts and transitions.
  • Wearable Energy Expenditure measurement;
    • Determining energy expenditure based on accelerometry signals.
  • Research Data Management Tooling;
    • Combining / integrating physical activity and indirect calorimeter data.
    • Combining physical activity, continuous blood glucose and continuous blood pressure measurements.
    • Combining indirect calorimetry, physical activity, CGM, CBP and nitrogen excretion data sets
  • Miss Activity eHealth app;
    • Cloud-based calculation of energy expenditure based on Predictive Equations (PE), Physical Activity Levels and Resting Metabolic Rate.
    • Integrate MOX Miss Activity wireless physical activity monitor in Food Friend platform to determine energy requirements.

New software for data processing of MOX accelerometry signals and physical activity report generation


The goal of this work package is to create and implement a software package for the operation of the MOX accelerometers. With this software, users should be able to configure the device, download the data, do a post processing analysis and generate a relevant report. The software needs to be modular, flexible and easy to support.

The software will have the following key features:

  • Being able to configure the MOX device and download the data after the measurement.
  • Analyze the raw accelerometer into relevant outcome measures.
  • Exporting the data in several formats after measurement.
  • Generating summary reports for participants
  • Generating editable reports for researchers with higher amount of detail
  • Storage of the raw data and the analyzed data according to the FAIR principle (Findable, Accessible, Interoperable, Reusable) and GDPR regulations.

Determining energy expenditure based on accelerometry signals


Acquire a combined data set of accelerometry, indirect calorimetry and heart rate data to update the activity classification algorithm and validate thresholds for the discrimination of LPA, MPA and VPA for healthy adults. This report will be updated with the combined insights of the different work packages. Once the METC protocol is approved data collection can start. A synchronized database from all different data sources will be the result of this task.


The primary analysis is to calculate the activity intensity using SMA and establish cut points for the discrimination between LPA, MPA and VPA. Secondary goals will be determined based on the research questions from the different work packages.

Study design

The goal is to collect data from healthy 50 participants (25 male, 25 females). The data collection will take place in the Metabolic Research Unit Maastricht. The participants will perform different activities of daily living such as deskwork (sitting and standing), sitting, standing and walking at different intensities. In addition, a resting metabolic rate measurement will be performed.

The following data will be collected:

  • Accelerometer data from the upper leg using the MOX1
  • Accelerometer data from the wrist using the MOX1
  • Accelerometer data from the hip using the MOX1
  • Accelerometer data from the chest using the MOX3
  • Lead II ECG data from the MOX3
  • Indirect calorimetry data using the Omnical and Basic Room Calorimeter
  • Movement intensity using a radar system

Based on the study and the goals of other work packages more data sources may be added. During the protocol, the participants will be video recorded in order to have a gold standard for post processing. With the output of data analysis, the activity classification algorithm can be further developed and validated. The primary outcome measures of the activity classification will be:

  • Sedentary Time
  • Standing Time
  • Dynamic Time
  • Minutes in LPA
  • Minutes in MPA
  • Minutes in VPA
  • Minutes in MVPA

Secondary outcome measures are sit to stand transitions, sedentary and dynamic bouts. The final algorithm will be implemented in the new MOX platform.

Physical activity measurement in room calorimeter studies

Physical activity is an important parameter to measure when performing and reporting whole body indirect calorimeter studies. This work package aims to characterize available physical activity assessment solutions and explore its integration possibilities in whole body room calorimeter study workflow.

In many calorimeter studies, a “run‐in” period is included in which exercise and energy intake are prescribed or controlled. Although not commonly practiced, a consistent sleep schedule may also be prescribed prior to calorimeter studies. The purpose of this run‐in phase is to minimize the impact of variations in these parameters on energy stores, EE, and substrate oxidation. Any prestudy restrictions (e.g., diet, physical activity, sleep, etc.) should be reported, along with the prescribed duration of these restrictions and how they were verified. Measures of physical activity can be used to report compliance with the pre-study protocol. Furthermore, recently it was recommended to always report PA related outcome measures in room calorimeter studies.

More advanced users of indirect calorimetry systems are aware of the possibilities of the combination with accelerometry. For example, sleeping metabolic rate is defined as the average EE during the three consecutive hours in the nighttime period with the lowest observed PA. To determine these three hours in a reliable manner a measure of body movement with accelerometry or an in-room motion sensor system is needed.


The first goal of this work package is to combine the MOX1 outcome measures with EE outcome measures.

The second goal of this work package is the combination of raw activity intensity and posture analysis with EE outcome measures. This combined dataset can be used for further analysis. An interesting secondary aim of this goal is the comparison of accelerometers with in-room motion sensor systems. For development purposes, it is needed that raw accelerometer data can be combined with raw EE data.

Combining physical activity, continuous blood glucose and continuous blood pressure measurements


The goal is to define which sensors and output parameters are of interest and how the sensors and data streams can be integrated in a merged file. Creating a user interface design for the software tooling, creating a design of a modular software tool to facilitate the use cases and develop the prototype including protocols for the data management technology.

The challenges are:

  • How to structure and organize the data in such a way to make it suitable for further analysis.
  • How should the raw data be structured and labelled to allow integration of different sensor modalities?
  • Defining format and storage of the codebook
  • How to present the assembled raw data set
  • How to ensure anonymization / pseudonymization
  • Which data cleaning procedures need to be carried out to make the data suitable for reporting and calculation? (time selection, removing time series, filtering, visual inspection etc.)
  • Calculated data / feature extraction: List of output parameters, calculations that need to be performed to achieve output parameters and data formats

Combining indirect calorimetry, physical activity, CGM, CBP and nitrogen excretion


The goal is to combine all data sources from the previous work packages. The software tools need to integrate all those parameters in one dataset.

Cloud-based calculation of energy expenditure based on Predictive Equations, Activity Level and Resting metabolic rate. Integrate MOX Miss Activity in Food Friend platform to determine energy requirements.


The goal of this work package is the development of a API / online processing interface that calculates energy requirements from a set of available data sources. The demonstration will be performed by integrating the MOX Miss Activity and EE-API with the Medrecord platform, taking data from both sources.