Contact · Newsletter · DEUTSCH
The interaction of the 6-axis inertial sensor (inertial sensor) MLS/160A with the machine learning modules of the Remote Maintenance Gateway RMG/941 results in a so-called soft sensor or virtual sensor, which uses the specific vibration characteristics of engines and machines as a functional principle.
Since the typical vibration characteristics are condition-dependent, such a soft sensor can be used for condition-based monitoring and predictive maintenance applications. The entire real-time data analysis can be performed directly in the RMG/941 without a permanent cloud connection.
Alternatively, a feature data vector acquired by the MLS/160A can be periodically transferred to the cloud and further processed there by a machine learning algorithm.
This enables condition-based monitoring based on real-time vibration data for practically any machine or plant.
The workflow of a Machine Learning (ML) based condition monitoring application consists of two phases. In a training phase, historical data with feature vectors are first collected from the sensors belonging to a specific application in a text file (CSV file) and then used to model a suitable ML algorithm.
In the subsequent inference phase, a single feature vector with real-time sensor data is then analyzed using the mathematical model by means of supervised learning and the respective operating state is classified.
The RMG/941 is delivered with a Python3 runtime environment with numerous data science libraries offering various ML functions up to neural networks.
The data streaming mode of the MLS/160A can be used both for the acquisition of history data and for the individual feature vectors.
The data of the 6-axis inertial sensor is sufficient for machines with low rotational speed to detect different states with high accuracy via ML-based real-time data analysis.
PyDSlog ("Python Data Stream logger") is a preconfigured software for data acquisition, which can be used to easily generate the feature vectors for modeling.
This allows an edge solution for condition monitoring with RMG/941 and MLS/160A to be implemented within a very short time.
With the PyDSlog Docker, we implement the data quality requirements of a machine learning application as precisely as possible and create standard interfaces (such as MQTT) to other applications.
The Docker can be used on SSV gateways as well as on other suitable platforms in the Edge and it does not only support sensor configuration, but is also used for secure remote update of the MLS/160A firmware.
You can find the PyDSlog-Docker on GitHub.
If you look at a typical controller solution in automation from the perspective of IT security, many solutions would actually have to be shut down immediately. The main reason is usually the lack of possibilities for software updates. In most cases, patches do not even exist, although some controllers have long known weaknesses.
Secure Device Updates (SDU) solve these problems and also offer the possibility to distribute new functions to the users. If a component, machine or system is to be supplied with software and configuration updates via an IoT connection, IT security must be taken into account in addition to system security.
Given the current state of the art, this requires a public key infrastructure (PKI) for digital signatures with private and public keys, certificates, revocation lists, etc., in order to at least guarantee the authenticity and integrity of the update. All necessary components are included in SDU.
With regard to firmware updates, the MLS/160A soft sensor is supported by both the SDU development tools and the RMG/941 SDU app.
I have 150 system controllers with MLS/160A as monitoring sensor in operation worldwide. There I would like to be able to maintain the sensor firmware via remote update if necessary.
We have published an extensive example of a firmware update of the MLS/160A via SDU and the RMG/941. The complete example of the firmware update including all code files and a detailed description can be found on GitHub:
| Processor | |
|---|---|
| Manufacturer / Type | STM32 Microcontroller with ARM 32-bit Cortex-M3 Core |
| Clock speed | max. 72 MHz |
| Memory | |
| RAM | 20 KB SRAM |
| Flash | 64 KB |
| Interfaces | |
| Serial I/Os | 1x RS485 half-duplex, 115200 Baud |
| JTAG/Debug | 1x 18-pin connector (internal) |
| Bosch BMI160 6-axis inertial sensor | |
| 3-axis accelerometer | Max. sampling rate: 1600 Hz Max. analog bandwidth: 684 Hz for x- and y-axis; 353 Hz for z-axis Measuring range: ±2 g (g = 9.81 m/s²) Resolution: signed 16-bit integer |
| 3-axis gyroscope | Max. sampling rate: 1600 Hz Max. analog bandwidth: 523.9 Hz Measuring range: ±2000°/s Resolution: signed 16-bit integer |
| Bosch BME280 environmental sensor | |
| Humidity | Operating range 0 .. 100% rel. humidity |
| Air pressure | Operating range 300 .. 1100 hPa |
| Temperature | Operating range -40 .. 85 °C |
| Software | |
| Operating system | RIOT 2019.04 |
| Other | SDU Device Driver |
| Protocols | RS485 for transmission of measured values in special data streaming mode |
| Configuration | Via datagrams |
| Miscellanous | Firmware update via Secure Device Update app (A/B boot) |
| Displays / control elements | |
| LEDs | 1x Power/system status |
| Electrical characteristics | |
| Supply voltage | 12 .. 24 VDC via external power supply unit |
| Power consumption | typ. 20 mA @12 V / typ. 10 mA @24 V |
| Mechanical characteristics | |
| Protection class | IP40/IP65 |
| Mass | < 50 g |
| Dimensions | 91.2 mm x 43.2 mm x 26 mm (without cable) |
| Operating temperature | -20 .. 70 °C |
| Standards and certificates | |
| EMC | CE |
| Environmental standards | RoHS, WEEE |
SSV SOFTWARE SYSTEMS
Dünenweg 5
30419 Hannover
Phone: +49(0)511 · 40 000-0
Fax: +49(0)511 · 40 000-40
© 2025 SSV SOFTWARE SYSTEMS GmbH. All rights reserved.
ISO 9001:2015
