No matter how fast or precisely they move, industrial robots only adapt as quickly as people can program them. Modified product specs, process improvements, customized orders—small changes like these add up to big costs whenever human intervention is required.
Artificial intelligence (AI) will revolutionize industrial automation for this reason. But it’s only possible with high-capacity, ultra-low-latency technology that can connect AI-scale inference processing to the factory floor. That’s where Intel’s technology leadership in 5G and edge computing are bringing the smart factory to life.
Faster Networks for Flexible Assembly Lines
Reinventing the digital bridge between servers and industrial IoT sensors will empower entirely new business models. By 2025, Tractica estimates the market for AI in manufacturing will exceed $13 billion. The same analysis forecasts 24 percent annual growth in this market, thanks to:
- Smaller capital expenses from wirelessly connected robots that require fewer fixtures that can complete a wider range of tasks
- Lower operating costs, as AI reduces the need for manual reprogramming
- Flexible factories where machines can be moved, reconfigured and repurposed in a fraction of the time it takes to rebuild a traditional assembly line
AI platforms in the cloud and sensor-laden equipment in the factory are growing more sophisticated by the year, demanding more from the networks. 5G brings the promise of enhanced bandwidth, sub 1-millisecond latency, 1,000 times the capacity of legacy wireless technology and massive machine to machine communication capability, unlocking the full potential of IoT in industrial production.
Edge computing plays a critical role in the smart factory. Moving compute closer to the data sources enables control and feedback into the AI algorithms with higher speed, lower latency and jitter. Combined with 5G, edge computing opens the door for capabilities not possible before adding an inherent layer of security and analytics with cellular protocols, and by storing data and proprietary content secrets locally.
A New Vision for Manufacturing
People often use the future tense when talking about 5G and AI-enabled robots, but in Stockholm, you can explore their use cases today. Intel’s 5G and IoT Innovation Center has demonstrated advances in intelligent vision that uses video analytics at the edge to give AI full control of an industrial machine.
Unibap’s Intelligent Vision System can direct the movement of any industrial robot. It uses a convolutional neural network to analyze video of the machine’s work and make operational decisions. This AI application leverages an 18-core Xeon™ edge server with the Intel Distribution of the OpenVINO™ toolkit to optimize learning and inference. At the same time, 5G and edge computing provide the ultra-low latency, high reliability and scalable compute resources required for implementation.
GE Digital, another project partner, analyzes the real-time production data at the edge and transmits settings to the equipment using its Manufacturing Execution System. The system also passes information that doesn’t contribute to real-time decisions to the GE Predix Cloud for deeper analysis and benchmarking.
The machine learning, edge computing, and 5G connectivity add up to a system that can rapidly reprogram a robot’s movements without human intervention. These combined capabilities allow robots and other devices to switch between small, highly customized batches during a production run. Learnings from a robot’s actions refine the machine’s performance over time while sending operators predictive alerts about maintenance issues. Just as impressive, the vision system can perform quality control—previously impossible for robots because industrial equipment lacked the capacity to “see” defects in the way a human can.
Exploring Innovation in 5G, IoT & The Edge
Intel opened the 5G and IoT Innovation Center two years ago to show the commercial potential of ultra-reliable low-latency communication from the client to the cloud. Early demonstrations explored smart buildings and the Skydome virtual reality (VR)/augmented reality (AR) system that gives first responders enhanced situational awareness during emergencies. Future industry partnerships will touch on autonomous driving, gaming, and immersive VR video streaming with eye tracking.
What all these projects share is the potential to take data-intensive applications beyond the limits of yesterday’s networks. In industrial automation, 5G opens the door of innovation, enabling the power of AI to proliferate, moving insights seamlessly from the data center to the point of highest value across the network. Intel powers this distributed intelligence, cloudifying the entire network by providing compute, storage, and analytics capability across the network, advancing manufacturer capabilities, efficiency and profit potential further than ever before.