)
)
)
)
)
)
)
How can we help you?
)
Marija Mijic
HR & Office Allrounder
T +43 660 383 95 46
office@craftworks.at
craftworks GmbH
Schottenfeldgasse 20/6a
1070 Vienna
Sensor Data for
Automatic Detection
Sensor-based automatic location, movement and state detection of concrete frameworks on building sites.
:quality(80))
Success Story
reading time: 3 min
Mechanical Engineering
Material Industries
Anomaly Detection
Industrial Software
80%
Accurate detection
Award Winning
Innovative solution
Self Learning
An annotation tool with self learning capabilities
Context
Our customer created a sensor that can be attached to concrete frameworks to monitor the construction process. The device includes an accelerometer and an additional sensor to measure the curation of the concrete. craftworks was assigned with the task to make the sensor data useful and give full insight in the progress of building works and the current state of the building site.
Challenge
To detect characteristic stages during the construction process, based on the accelerometer data of the devices, which are applied to the frameworks. Upon detection, our client needed answers for a series of questions: Is a framework being lifted by a crane? Is its position horizontal or vertical? Are there workers on the framework? Have parts been attached to the framework?
Assignment
Our task was to collect accelerometer data on-site, log all the significant stages through labelling and train a machine learning model to automatically detect the stages.
Solution
We developed the code in Python, then ported the machine learning model to C code in order to run it on the microcontroller of the customer's device. We also used Pandas and Scikit-learn. For labelling, we used our efficient in-house build annotation tool, which has self-learning capabilities.
craftworks' innovative solution, integrating the customer's sensor data, automates concrete framework monitoring with 80% accuracy. Our approach enhances project management, optimizing building sites efficiency.
:quality(80))