Drones make stock-taking a lot faster and more efficient. During the Digital Logistics Congress, Avular showed that drones can be used for many more applications in warehousing, from early detection of fire to tracking uninvited guests. "The possibilities in warehousing are endless. We just need to train the AI system for each new application to be able to translate the camera images into meaningful information."
The national media are full of stories about drones being used during the wars in Ukraine and the Middle East. These are drones that fly on fuel and usually do not need to return. "That is different with our drones that we deploy in warehousing. Those do have to return, but they have one big disadvantage: the battery," Berend-Jan Taken, CEO of Avular Logistics Solutions, told me. "To be able to fly for twenty minutes, a drone has to spend forty minutes on the charger. If you want to use a drone for long periods and intensively, you quickly have a capacity problem."
Avular designs and manufactures robots and drones, including for applications in warehousing. As a solution to the capacity problem, the Eindhoven-based company connected its drone via a cable to an autonomous mobile robot. Together, they drive independently through the warehouse to take stock. To do so, the drone is equipped with a camera, which takes images of all stored pallets. As the battery is on the mobile robot, the drone can now stay airborne for six hours at a time.
Sixty per cent savings
Combined with the mobile robot, the drone can capture some 8,500 storage locations on video in six hours. The images are then analysed with the help of artificial intelligence (AI) and converted into useful information. That information is then compared with information in the warehouse management system (WMS). "If the WMS says that a storage location is empty, while our images show that there is indeed and pallet there, you have an inventory difference. We record these stock differences and compile them into a report," Taken explains.
Those who want to inventory their stock without a drone will have to send employees out to count all the pallets. These should not only look, but also write down what they see. That information must then be processed again and compared with the stock information in the WMS. It is estimated that companies active in warehousing can save around 60 per cent on the cost of manual stock counting with a drone.
Other applications in warehousing
Using drones for inventory counting is not new. Logistics service providers in particular like to use this technology to give their clients reliable information on stock levels. Increasingly, they are getting ideas for other applications in warehousing. "One of our customers wants to use the images to check whether its employees attach pallet labels in the right place. These should be at the front of the pallet, preferably in the middle. If a forklift driver has to scan the pallet, he doesn't have to look for the label for a second."
One possibility is to equip the drone with an infrared camera. Avular uses that camera in warehouses where chemicals or lithium batteries are stored. With it, they can detect early temperature rises. These could indicate a fire or other dangerous situation. Tasks: "This kind of application also requires the deployment of AI to analyse the images and translate them into meaningful information. Using an employee to look at those images in real-time is obviously not an option."
With AI, it is possible to count boxes in addition to pallets. "Based on the images, AI can see what the stacking pattern of a pallet looks like. If the drone then captures a broken pallet, AI can calculate how many boxes are gone. Assuming that the order pickers have neatly started taking out the boxes at the front," says Taken. "Similarly, we have beer brewers using our drone to count how many beer crates they have outside."
Positioning in warehousing
Chief technology officer Erik van Meijl explained that Avular develops everything itself, both hardware and software. He showed how a drone uses a 3D LiDAR sensor to scan the environment and measure the distance to walls and objects. That measurement data is used by Avular's platform to create a 3D map of the warehouse. The drone uses that map to find its way through the warehouse. "We can also use that data to determine the exact position of a pallet or measure its volume," he says.
Van Meijl mentioned mounting a sound sensor so that the drone has ears as well as eyes. These ears make it possible to inspect technical installations: defects and leaks can be determined with an accuracy of a few millimetres. "Also think about detecting movements in the warehouse. You can see when an uninvited guest walks into the warehouse and where he is going," Van Meijl explains.
Taken adds: "We have now received a request from a company that stores colourless, transparent chemicals. This company wants to use our images to detect leaks. We are now investigating whether we can train our AI system so that it can recognise leaks of transparent liquids. The sky is the limit, in terms of applications in warehousing."
