Bringing more efficiency and safety to warehouse logistics
Global trade has significantly risen in the past decades. The growth of e-commerce has been firm and has recently accelerated due to the strong impetus of the pandemic. This led to an immense growth in logistics. The global logistics market is estimated to grow to from USD 10.32 Billion in 2017 to USD 12.68 Billion by 2023 as predicted by researchandmarkets. In particular, one will see a growth in the amounts of goods that are transported to and from warehouses, logistic centres, distribution centres, etc.
As only 0.2% of current warehouses is automated, this means that for the un- and offloading happens with driver-operated forklifts.
The global forklift market size valued at $49.6 billion in 2019 with about $451 million towards automation and is expected to grow at an annual growth rate (CAGR) of 12.1% from 2020 to 2027 according to grandviewresearch and fortunebusinessinsights.
The forklift operation is a physical job that requires certification to operate; making it hard to keep staffed with trained staff. The current boom in e-commerce retail over the past year has significantly inflated the demand of forklift operations in warehouses causing the bottleneck for movements of products. This opens the opportunity for automation. While on the profitability side, automation can deliver round-the-clock operations covering several shifts and multiple locations. The self-controlling forklifts can coordinate in terms of routing and speed. The key benefits include fatigue free workspaces, substantially less damage to shelving, loads, pallets, and other trucks, with more accurate navigation.
There is a great need for this technology in this area. How to get from today’s manually operated situation fully self-operated forklifts? The first step is equipping forklifts with the right suite of sensors mounted at the right locations on the forklift that enable to train and run the autonomous software. Product-wise, it’s easier to enable a full automation in forklifts. Today’s forklifts have enough room and power for data processing and have digital controls for computer-steered actuations.
The way Teraki enables automation for forklifts is to start with remote-control and gradually but quickly go first to semi-autonomous and later full-autonomous software (AI-models). Teraki takes care of recognition of surroundings and objects, training of tasks and how to handle events, path planning and triggering of actuators.
To fully enable the automation into the forklifts, we start with remote operation, like cars and delivery robots as described in our previous blog. The forklift (or other heavy-duty vehicle) will be equipped with essential sensors - typically cameras, or camera and Lidar - and connectivity. Subsequently, the Teraki software will be deployed in the vehicle to extract the relevant information from the raw sensor data. This results in 4x – 5x less data which is key for effective transmission at low latency, without any qualitative impact on what the remote-control operator sees on the screen. In addition, network bonding is applied. This part of the software dynamically seeks the highest available bandwidth between multiple wireless networks (cellular and/or WIFI) and continuously uses and changes between the connections with best conditions.
The object detection and bonding solutions combined lead to best-in-class operational performance when it comes to low latency (sub 50ms) and number of disconnections (typically: none). One can even operate in areas with 300Kbps of bandwidth availability.
The forklift itself will be equipped with advanced safety assist systems such as collision avoidance, auto stop features developed for automotive driver assist standards.
AV-stack: delivering autonomy
For Teraki the remote operation is a first step towards full autonomy. While the forklifts are being remotely operated, in the meanwhile the self-driving AI-models are trained: continuously learning takes place from the actions and tasks that are performed and the environment in which they are performed. This learning process is an essential step to deliver partial (and finally full) autonomy.
The most important technology applied in the remote driving process is the Region of Interest (ROI) and Time of Interest (TOI) detectors. The region of interest in the camera of the forklift recognizes and captures the relevant and essential objects (such as warehouse line markings, objects, people, etc.) in video stream and sets a high-quality factor for the detected ROI, this drastically helps keep the transmitted file size low (75% - 80% additional reduction) for the high streaming quality with low latency to ensure safe and real-time operation without disruptions. This goes without any impact on the visual quality perceived by the human remote driver (measured in e.g. VMAF).
The TOI detector selects the important events from the operation of the forklift such as a turn, detected objects, overtake situation, crossroads, etc. The extracted TOI from the remote-driving operation significantly reduces the training time for autonomous software (30x less training time) and brings that to the required 99%+ accuracy to be deployed safely in the operation.
The two big benefits of remote control and automation are safety and cost reduction.
The remote operation gets the tasks of forklifts done without having the human in a remote site decreasing the risks of injuries, health exposure (sound, air, vibrations, etc.) and lowering health insurance costs.
Moreover, the ‘disconnect’ between physical presence of the operator and the forklift enables companies to allocate drivers between warehouses in seconds. Operators can be distributed over multiple (geographically spread) projects. If there is peak demand any of the warehouses, operators in (one of) the remote driving hub(s) can be allocated to that task in seconds. Contrast this with the current situation of requiring a pool of operators physically at each warehouse to cope with peak demand in that specific warehouse. Requiring a buffer for each location. Instead, one can allocate a central pool of operators and ‘switch them from Paris to Berlin to Amsterdam’ in seconds. Overall, less 20% operators are required. In addition, companies and can work outside “office hours” as one can apply “follow the sun” philosophy.
One can see three stages of efficiency improvements.
By centralizing remote drivers in one or more central ‘remote driver hubs’ allocating staff to forklifts jobs throughout the warehouses will be about 20% more efficient. Agile handling of peak demand becomes possible and there is no need for local ‘buffers’ of forklift drivers per location.
By automating the easy, repetitive tasks of the forklifts, one remote operator can handle more forklifts. The operator only needs to operate the forklift for the loading / unloading and other ‘difficult’ situations. The movements from A to B are done autonomously. This delivers a “1 : N” ratio of “driver : forklifts”. Depending on the situation, this can go up to 1:5.
The third and final step is full autonomy where the remote drivers can be taken out of the loop and operations are run completely autonomous.
The remote operation provides a safer working environment, removing the physical risks, eliminating operation fatigue, reducing the costs of safety labour regulations and insurance. Finally, it offers handicapped people the opportunity to do the job. A great way to involve this group of people in the labour market to a function they could not have done before.
With the stong growth in (e-commerce driven) global logistics and with the skilled labour shortage, the need for automated operation of forklifts has never been more significant. Teraki helps implementing the automations in steps and offer a short roadmap to full autonomy with our safe and AI-powered software. Teraki’s AV-stack allows for an easy & manageable start and can easily be scaled. We help customers to improve efficiency, costs and agility in their daily loading operations throughout all their warehouses and logistical centres.
Get in touch
Feel free to contact us at firstname.lastname@example.org for a first conversation what we could mean for your business.