AGVs - the 'pollinator bees' of flexible manufacturing
02 October 2020
Modern manufacturing is becoming more fluid: Production, assembly and materials handling processes are increasingly flexible, modular and self-organising. Neil Sandhu looks at how AGVs fit into this landscape
Conventional models with fixed, single-product, linear mass-production are heading for the history books. In their place, Industry 4.0 manufacturing and logistics operations are highly adaptable, switching seamlessly between smaller batches made to increasingly-customised specifications.
Most operators don’t have the luxury to start from scratch and build from a Smart Factory blueprint to accommodate these new, highly-dynamic processes. Instead, legacy fixed assets must be adapted to modern production processes. Decommissioning or moving existing plant and machinery would just be too costly. So, automated, driverless mobile machines of all kinds are becoming an increasingly popular and cost-effective addition to shop floors and warehouses, because they help to adapt the fixed infrastructure to modern production processes.
Joining the dots
Automated transport machines of all kinds are replacing load carrying by fixed conveyors and rails. They can be like pollinator bees, flitting between machine cells and stores to connect processes together. Whether automated carts, pallet trucks, forklifts or mobile robots, they join the production dots.
Once goods leave a conventional conveyor system, their location can become unclear. Increased digitisation also offers the opportunity to create a transparent overview with a data trail of the entire production and logistics chain.
As a result, the travel paths of load carriers can be optimised and adapted in real time, scheduling can be more responsive, and a material flow can be planned and controlled according to consumption. It’s even more of an advantage if the vehicles can also be tracked using GPS, and the data collected for cloud-level monitoring and management. The possibility for ‘black hole’ interruptions between known, fixed points in the production trail is eliminated, and on-time deliveries can be optimised.
The starting point for this more automated and adaptive future is that the vehicles must know where they are in time and space. Indoor localisation is the process of tracking an object, person, or equipment through a warehouse or logistics centre. Localisation technology is one of the key materials handling technologies needed to achieve the type of networked production and logistics needed for Industry 4.0.
There are a range of AGV localisation technologies that can be fitted onboard AGVs - ultra wide-band tags, contour mapping, line guidance sensors, infrastructure sensors - to record either their own position and/or the position of the target objects they are sensing. And, with connected sensors and software analysis tools, large amounts of data can now also be processed for display in higher-level factory and logistics systems.
Developments in laser scanning and LiDAR (Light Detection and Ranging) technology have been critical to improvements in AGV navigation. Typically, LiDAR sensors have enabled localisation by using reflectors, or by recognising the contours of their environment as the vehicle travels through it, using a pre-taught map. Now, LiDAR technology is being used on AGVs, service robots, shuttles and unmanned forklifts to set vehicles completely free from the slavery of fixed paths.
LiDAR sensors use “time of flight” measurement to calculate the precise distance and direction of all kinds of contours detected ahead of, or around, the vehicle. The measurement data can be used to identify walls, gates, pillars or shelves and map the vehicle’s operational environment.
So called Localisation-on-Contour systems can not only learn and recognise the physical contours of a factory or warehouse interior, but are adaptable whenever the surroundings are altered. The SICK LiDAR-LOC, for example, puts an end to costly installation and maintenance of reflectors, special paints, tracks, magnetic strips or coloured tapes.
Setting up is as simple as ‘teaching’ the on-board LiDAR scanner prominent contour features, such as walls, large static machinery, racking or bays, as the AGV is driven manually around its working environment. Then, this data is used by the service team to create a precise reference map before being uploaded on a Sensor Integration Machine and easily commissioned on-site.
During operation, 2D data from the LiDAR sensor matches its position with the map stored in the SIM and also mounted on the AGV. As the data is processed, an algorithm continually compares the distances retrieved from the scan data with the map to provide position and orientation information to the AGV controller.
Use existing scanners
To save duplicating scanners, contour-based navigation can be retrofitted to a vehicle by using scanners that are already installed, most likely as part of the safety system. As a software-only system, the SICK LiDAR-LOC allows you to set up localisation using other SICK LiDAR sensors, including safety laser scanners. This can be especially advantageous when designing for smaller and low-to-the ground mobile platforms where compact designs are a critical consideration.
With all this increased autonomy, engineers might reasonably expect extra challenges in integrating both navigation and safety technology into mobile vehicles, and potentially both increased hardware costs and software development time.
However, LOC systems using one or more SICK LiDAR sensors can be set up with minimal programming time. As a result, software development, installation, hardware and maintenance costs can all be dramatically reduced.
One challenge associated with autonomous vehicles using magnetic or optical-guide tape navigation is a loss of tracking guidance that leads to the potential for unplanned vehicle stops and fleet jams.
A hybrid version of the SICK LiDAR-LOC is also available to act as a supplementary system to magnetic or optical-guide tape navigation. Using the software from SICK, AGVs, carts, and service robots can continue to remain on track even if the tape happens to be missing. The system can use the 2D measurement data of the surroundings acquired, for example by SICK safety laser scanners, which are often already present onboard the vehicle.
The system switches seamlessly from tape guidance to contour-based navigation whenever the tape can no longer be detected because it is missing or damaged. The system can be integrated into a new machine design or retrofitted to an existing vehicle.
It travels along with the vehicle as a supplementary system and activates itself automatically only when the tape can no longer be detected. The switching is unnoticeable to the observer as the vehicle continues to travel reliably along its route without any hesitation.
The track is taught in during the reference run of the vehicle and mapped electronically. If the track on the ground is lost, the vehicle controller switches to the backup measuring system which accurately determines the vehicle position. When the tape is detected again, the track guidance system takes over the vehicle navigation as before.
As sensors become more intelligent and programmable, integrating systems into mobile vehicles is becoming more straightforward, whether you prefer to harvest raw data for your own development, or to take advantage of ready-to-go software solutions. Increasingly, it will also be possible to gather and track diagnostic and condition data for service and maintenance.
Neil Sandhu is SICK’s UK product manager for imaging, measurement and ranging