Abstract
Manufacturing systems for individualized production require workflows depending on individual objects. Machine learning (ML) offers the possibility to classify different objects by training a neural network. Depending on the output values of the network, decisions for the following production step can then be controlled. The question arises whether it is possible to execute the neural network in real time in coordination with the machine and motion control tasks. In this paper, this question is investigated using a programmable logic controller (PLC) runtime environment on a standard industrial PC. The execution times of different neural network implementation methods are measured and compared. The fastest neural network requires an average execution time of only 54 µs. The characteristics of the different methods with respect to training and implementing the neural networks in the controller are also discussed.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
It should be emphasized that results shown above apply to an industrial controller based on an Intel multicore processor. There are also possibilities to run the neural networks on separate controllers which adds additional interfaces and latencies. For example, Siemens offers the TM-NPU expansion module [16] for the S7-1500, on which neural networks can be executed.
References
Koren, Y.: The local factory of the future for producing individualized products. In: The Bridge, National Academy of Engineering, vol. 51, no. 1, pp. 20–26, 1 April 2021
Hornberg, A.: Handbook of Machine and Computer Vision. Wiley-VHC (2017)
Wani, M.A., et al.: Advances in Deep Learning. Springer (2020). https://doi.org/10.1007/978-981-13-6794-6
Beckhoff, TF7xxx | TwinCAT 3 Vision. Accessed on 24 Oct 2021. [Online]. Available: https://www.beckhoff.com/en-us/products/automation/twincat/tfxxxx-twincat-3-functions/tf7xxx-tc3-vision
Keras, Accessed on 21 Oct 2021. [Online]. Available: https://keras.io/
MathWorks, MATLAB. Accessed on 24 Oct 2021. [Online]. Available: https://de.mathworks.com/products/matlab.html
ONNX Supported Tools. Accessed on 21 Oct 2021. [online]. Available: http://onnx.ai/supported-tools
Li, M., et al.: The deep learning compiler: a comprehensive survey. In: IEEE Transactions on Parallel and Distributed Systems, vol. 32, no. 3, pp. 708–727, 1 Mar 2021. https://doi.org/10.1109/TPDS.2020.3030548
University of Applied Sciences Kiel, et al.: Machine learning for image recognition in the real-time context of industrial control systems. Accessed on Oct 21 2021. [Online]. Available: https://www.youtube.com/watch?v=KmPGQ2w_Peg&t=3s
University of Applied Sciences Kiel, et al.: Real-time object recognition with neural networks for industrial control systems. Accessed on 21 Oct 2021. [Online]. Available: https://www.youtube.com/watch?v=DXfj88JkX5YY
Le Cun, Y., et al. The MNIST database of handwritten digits. Accessed on 21 Oct 2021. [Online]. Available: http://yann.lecun.com/exdb/mnist/
Smeschke: Four shapes. Accessed on 21 Oct 2021. [Online]. Available: https://www.kaggle.com/smeschke/four-shapes
Google Colaboratory. Accessed on 21 Oct 2021. [online]. Available: https://colab.research.google.com/
Beckhoff, TwinCAT 3 Engineering. Accessed on 21 Oct 2021. [Online]. Available: https://www.beckhoff.com/en-us/products/automation/twincat/te1xxx-twincat-3-engineering/te1000.html
Beckhoff, TF3810 | TwinCAT 3 Neural Network Inference Engine. Accessed on 21 Oct 2021. [Online]. Available: https://www.beckhoff.com/en-us/products/automation/twincat/tfxxxx-twincat-3-functions/tf3xxx-tc3-measurement/tf3810.html
Siemens. SIMATIC S7-1500 TM NPU, Accessed on 21 Oct 2021. [Online]. Available: https://new.siemens.com/global/de/produkte/automatisierung/systeme/industrie/sps/simatic-s7-1500/simatic-s7-1500-tm-npu.html
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Wree, C., Raßmann, R., Daâs, J., Bause, F., Schönfeld, T. (2022). Real-Time Image Analysis with Neural Networks on Industrial Controllers for Individualized Production. In: Borangiu, T., Trentesaux, D., Leitão, P., Cardin, O., Joblot, L. (eds) Service Oriented, Holonic and Multi-agent Manufacturing Systems for Industry of the Future. SOHOMA 2021. Studies in Computational Intelligence, vol 1034. Springer, Cham. https://doi.org/10.1007/978-3-030-99108-1_34
Download citation
DOI: https://doi.org/10.1007/978-3-030-99108-1_34
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-99107-4
Online ISBN: 978-3-030-99108-1
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)