Four application areas for human-machine collaboration in the smart factory

In my last posts, I have written about the AI revolution in manufacturing and the need for human-machine cooperation in smart factories: Impactful performance boosts will only happen when human workers and AI-enabled smart machines work together.

AI automates machines, processes and empowers human workers to respond to unexpected situations. Thus, workers can make smart decisions based on the proposals of the AI. In smart factories, humans and AI closely collaborate and bring tremendous benefits to various functions.

The following use cases showcase the cooperation between humans and AI and focus on efficient human-machine interfaces to connect human workers with AI based systems.

Human-Machine collaboration in Assembly

Using robots is already state-of-the art in almost all industries. Especially for tasks that require heavy lifting or repetitive work, robots reduce costs and increase speed, thereby boosting productivity. With the rising need for individualised products and lot size 1, the complexity of production is increasing and more and more flexibility is needed. Rule-based robots with fixed programming schemes will fall short in coping with the complexity of production. They will be replaced with more flexible AI-based systems. Those systems are a combination of AI and flexible collaborative robots – or cobots. The systems will become self-optimised systems that adjust to changing production requirements in real time. They allow direct interaction between human workers and machines. For such collaboration, mobile user interfaces are needed to allow direct interaction between cobots and human workers.

VeoRobotics builds high-performance industrial robots that work collaboratively with human workers to enable much more flexible, productive, and efficient manufacturing work cells.

Human-Machine collaboration in Maintenance

AI can be used to reduce machine breakdowns and increase utilisation of production assets. As such, predictive maintenance algorithms avoid breakdowns by continuously analysing parts and replacing them before incidents occur. As a prerequisite, machine data and human workflow data need to be collected on a continuous basis, so that the AI can continuously learn from real-time production data. Many technology providers already offer systems that connect machinery with industrial internet of things backend applications. Such systems can be installed on premise or in private clouds.

oee.cloud offers a solution for the optimization of OEE. The AI is connected with minimal intrusive sensor hardware to collect data about overall equipment effectiveness. All collected data can be analyzed and if required, notifications can be sent to service technicians. For example, the Workerbase smartwatch allows to notify service technicians in case of incidents. The solution not only sends real-time alerts but can also be configured to show maintenance instructions. These instructions are based on insights of the AI, so that technicians always get individualized instructions based on context data, e.g. skill level, language, workflow progress etc.

Human-Machine collaboration in Quality assurance

AI systems can be used to identify quality issues as early as possible. High-performance cameras attached to machines utilize image-recognition technology. This helps to automatically recognize defects in products. As AI systems can learn continuously, they improve over time resulting in early identification of defects. Production lines can be stopped if a failure or deviation is detected, and operators can be alerted to minimize the impact of the deviation, thus reducing scrap and the need for rework.

Instrumental increases visibility into the assembly line by comparing photos of units and automatically analyzing batches of units for defects. The solution allows to track down defective units based on serial number, build date, or stage of assembly. By using the AI-based solution, issues can be discovered early in the production process, e.g. before moving into mass production. Human experts can then validate designs and fix the production setup without sacrificing quality or schedule.

The WORKERBASE smartwatch offers a lightweight wearable camera. Thus quality incidents can be captured from the shop-floor. Pictures are then directly send to a server system, ready to be analysed by AI bots.

Capture quality incidents directly from the shop-floor

Human-Machine collaboration in Logistics

Making multiple product variants and customer-tailored products results in increasing logistics activities in a factory. AI systems manage autonomous movement and efficient supply of material. For example, Autonomous guided vehicles (AGV) transport material from the warehouse to individual production cells. They use AI to sense obstacles and autonomously determine the optimal route. Thus, AGVs can cope with the increasing complexity that comes with low-volume/high-variance production.

Furthermore, machine learning algorithms will use logistics data such as inventory levels or flow of material to self-optimize warehouses. For example, low-demand parts can be automatically moved to remote areas of the warehouse while high-demand parts will be stored at nearby areas for faster access.

6 River systems offers AGVs for flexible warehouse automation. Their solution is a “cobot for logistics” following warehouse workers through their work zones to minimize walking and work more efficiently.

The WORKERBASE smartwatch supports all kinds of logistic tasks, e.g. interacting with AGVs, receiving picking instructions, sending transport requests or replenishment orders. The device allows to collect logistics process data, e.g. picking sequences which can be analyzed by an AI system to optimize the warehouse setup. By using language generation and processing, the device can give workers context-specific information, e.g. for pick-by-voice to handle picking and commissioning operations.