Whether raw data or pre-processed information: Neither data transmission nor bidirectional communication is possible without a suitable industrial interface. However, production and logistics concepts in the context of Industry 4.0 require more and more communication-capable sensors. The demand for IO-Link-capable solutions is therefore increasing noticeably, as they offer the possibility of creating full transparency and control right down to sensor level. In addition to pure detection or process data, sensors can also transmit information for process monitoring. This information is important for predictive maintenance or for process optimization in terms of reliable and robust detection. Set-up parameters can also be transmitted in this way.

But what can be done about the flood of data generated by communication-capable sensors at the lowest level? Sick relies on the concept of IO-Link-based 'Smart Sensor Solutions'. The core of this concept is data acquisition and its conversion into the application-relevant information actually required directly in the sensor. To this end, the sensors are equipped with intelligent automation functions, allowing self-contained subtasks to be completed faster and more efficiently than in the machine control system. The result:

• Condition monitoring already in the sensor enables active self-monitoring and therefore predictive maintenance.
• The process speed of the machine and its performance are increased. The sensor calculates the information required for the control process directly and then forwards it to the machine control system.
• Pre-processed information is transmitted to the control system rather than large, computationally intensive and time-consuming volumes of data. Data preparation in the control system is no longer necessary.
• The measured values determined independently by the sensor are more precise, as the fluctuation caused by the cyclical reading of the pulses into the controller is eliminated.

Increased control options

The transparency of the automation functions and the parameters of a sensor across all levels of the automation pyramid increases the control options for entire processes: This is because, in addition to the pure process data and application function information that is constantly available to the controller in real time, 'smart sensors' can provide further information for process monitoring and evaluation. As this information can be queried by the control system as required, it is also possible to analyze errors or monitor a sub-process.

More control also means being able to monitor and, if necessary, display target deviations within a complete process. Fork sensors from Sick, for example, detect labels and simultaneously provide information on the actual number of labels on a banderole. This allows deviations to be detected and reported to the label manufacturer.

2D and 3D vision solutions

In the field of 2D and 3D vision, it is also important to pre-process data at the point of origin. 2D and 3D vision solutions are suitable for applications involving inspection, measurement, localization or identification. The requirements for the systems are clearly defined: Capture the environment, analyze data immediately and deliver immediate results from which clear measures can be derived - even under the most difficult conditions. This means that in addition to reliable image capture, efficient data processing directly in the sensor is crucial.

Big data is already a reality for goods flows. Multiple sensor data with allocation to individual goods allows conclusions to be drawn about the quality of the logistics process worldwide. Clearly identifiable objects are assigned properties such as volume, weight and condition documentation with the help of sensors. Production companies, on the other hand, are required to prepare themselves for the new data structures. For example, simplified operation of system components, such as intuitive apps that can be downloaded directly into the machine, is proving helpful.

The '3vistor-P' 3D vision sensor: sensor, display and evaluation unit form a complete package.

© Sick

New challenges are constantly arising in the context of Industry 4.0: For example, the desire for flexible, customer-specific production - 'batch size 1' - results in a greater number of variants, but high machine availability and production efficiency must still be guaranteed. This is where 3D vision sensors come into their own: Because even if the size, height or shape of objects vary, the sensor reliably detects them and provides the necessary information. There is no need to re-parameterize a machine or system - for example, when changing products.

Below are three examples from Sick of 3D vision sensors based on Smart Sensor Solutions:

The 'TriSpector1000' 3D vision sensor is suitable as a stand-alone vision sensor for 3D inspections such as checking the contents of containers or quality control of consumer goods. An integrated image analysis simplifies parameterization; the sensor is also available with pre-calibrated 3D data on request. Intensity values are added as an overlay over the 3D data. This allows the sensor to check the presence and position of labels or printed patterns.

The 3D vision sensors of the '3vistor-P' and '3vistor-T' product families use 3D snapshot technologies. Based on stereoscopy, the '3vistor-P?CV' calculates the spatial and depth information of the objects in its vicinity in real time, even for stationary objects. The intelligent data evaluation makes the sensor suitable for driver assistance on heavy, off-road commercial vehicles in ports, mines or on construction sites. If an object is located in one of the two pre-configured alarm areas, the sensor triggers an audiovisual alarm via the display in critical situations. The sensor provides fully pre-processed information and data-reduced digital signals.

The 3vistor-T provides depth information for each pixel in real time based on time-of-flight measurement. This provides up to 30 three-dimensional images per second. The information actually required is calculated directly in the sensor. Thanks to the data reduction, the sensor is particularly suitable for vehicle applications such as collision warning, obstacle detection or navigation support.

Image processing and Industry 4.0

In future, the importance of multiple analysis with one vision sensor will grow. 3D position data for collaboration with robots and grippers will increase flexibility; reliability data parallel to reading results will flow into the decision-making processes. In addition to the direct results from image processing, such as good/bad decisions, intermediate results, probabilities and reliability data play a major role in higher-level processes. It is only through them that intelligent systems and automated process optimization are possible. Access to such data, for example via OPC-UA, is a prerequisite for the cyber-physical production systems of tomorrow.

Author:
Andreas Behrens is Head of Barcode/RFID/Vision at Sick in Waldkirch.