Report expects automated systems in logistics and fulfilment to explode over next 20 years

Report forecasts that the market for mobile robotics and drones in delivery and warehousing is likely to $290 Billion by 2040

Report forecasts that the market for mobile robotics and drones in delivery and warehousing is likely to $290 Billion by 2040 [Credit: IDTechEx]

A new report has estimated that the market for mobile robotics and drones in delivery and warehousing is likely to reach a staggering $81 and $290 Billion in 2030 and 2040, respectively. This growth will be uneven, with collaborative and autonomous robots within warehouses leading the way. 

Guided warehouse vehicles out, racking robots in

The Mobile Robots, Autonomous Vehicles, and Drones in Logistics, Warehousing, and Delivery 2020-2040 report expects the lack of adaptability in current Automated Guide Carts And Vehicles (AGC and AGV) and their infrastructure dependence will lead to poor growth expectations. The authors say that as they are neither fast nor efficient utilizers of space and do not lend themselves to human-robot collaborative workflows they may well decline in usage. The report “finds that their market will shrink by 50% in 2030 compared to the 2019 level.”

Instead, as the technology is evolving toward more autonomous and infrastructure-independent navigation goods-to-person automation within fulfillment centres will be a major growth area.

Special robot-only zones are created within warehouses in which these robot fleets move racks at high speeds to a manned picking station. The report notes that productivity gains are clear and proven. The technology results in space-efficient and compact warehouses. The pick-rates significantly improve and the headcount drops.

Many product design innovations helped enable this market. The hardware requires good acceleration and deacceleration to operate at high speeds with minimal spillage. The racks require special adaptations to steady the load during transit. The suspension system – which lifts and lowers the racks – require special design and engineering. The robots will include multiple motors, allowing them many degrees of freedom in movement. The navigation technology itself is not complex though as it is often fiducial based, mainly in the form of barcodes often printed at regular intervals on the ground.

The report notes that the key value-add technologies are on the software side though. This includes the entire stack from customized firmware sitting on the motor drivers all the way to the fleet and task management levels. The business model is also an important differentiator with many moving toward becoming full solution providers.

This is a fast-growing market space. The report authors say that the landscape was set on fire when Amazon acquired Kiva Systems for $775M in 2012, thereby leaving a gap on the market. Today, significant well-funded alternatives such as GeekPlus (389$M), GreyOrange (170$M), and HIK Vision ($6Bn revenue) have emerged, achieving promising and growing deployment figures. The number of start-ups has also increased, especially around the 2015-2017 period. Not all will be successful though, even if they offer a good enough technology. In particular, doubts over the long-term viability of some start-ups' business models and financial health act a barrier against long-term adoption by major end-users.

They forecast the annual unit sales to double within 6 years, with the global inflection point to arrive around 2024, beyond which point the pace of deployment will dramatically accelerate. The report forecasts that between 2020 and 2030, more than 1 million such robots will be sold accumulatively.

Towards autonomous indoor robots and vehicles

The outlook for classic AGV/AGCs does not look bright, however. A major reason is that the navigation technology is transitioning from automated to autonomous. The primary benefit is that the navigation becomes infrastructure-independent, allowing the workflow to be easily modified. The autonomous mobility also allows various modes of collaborative workflows between robots and humans, thus extending the utility of such autonomous mobile robots and vehicles to existing facilities.

The technology is enabled by better SLAM algorithms. The algorithms – based on different sensors including stereo camera and 2D lidars – are evolved enough to handle safe autonomous navigation within many structured indoor environments with a high degree of control and predictability. These robots are easy to install and to train.

The technology options, however, are still many, and choices have long-lasting strategic consequences. A common process is to use 2D lidars to develop a map of the facilities during the training phase, e.g., walking the robot around the facilities. The fixed reference objects will be marked during the set-up phase. This system is fairly simple. However, it does not deal well with dynamic and changing environments. Another approach is to use camera vision and deep learning to also identify and classify objects. This is computationally more complex, but will enable a more flexible system that can have more intelligent decision making in a complex and changing environment. It extends the utility of such vehicles to more indoor scenarios and allows a mobile robot that intelligently responds to its changing environment.

The business models are also various and evolving. Some are offering their technology as RaaS (robot as a service). The idea here is that the users do not need robotic expertise, do not require upfront capital, and need not worry about the risks of technological obsolescence and change. The RaaS model also fits into the operational budget of the users, further facilitating adoption. On the other hand, many follow a traditional model of equipment sales. Even these suppliers will also need to build a subscription platform into their business model to offer maintenance and upgrades, especially cloud-based software updates. In other words, even these models will involve a large element of service revenue.

There have also been some notable recent acquisitions. Amazon acquired a company focused on camera-based navigation, which would enable object detection and classification, and thus more intelligent navigation. Shopify acquired a firm with full-solution including the entire software stack. This company also had RGB camera-based processing technology. The acquisition was at a very high multiple, but will give Shopify control over a full strategic technology essential in its recently-announced drive to diversify into the fulfillment business.

Overall, the report forecasts that more than 200k robots could be sold within the 2020-2030 period (this figure includes those that can perform picking of regularly or irregularly shaped items). 

Forklifts and tuggers to also go autonomous?

According to the report, forklifts and tuggers are set to go on a journey from manually operation to autonomy, although this is still more than half a decade from widespread maturity. It says many have already developed, demonstrated, and deployed autonomous forklifts and trucks. 

The choice of navigation technology is varied. Some use RGB camera and RGB image processing technologies to navigate. This field was extremely challenging in the past. Recent, deep learning advances, however, have seen this field completely transformed over the past six to seven years. These strides now enable excellent localisation and 2D object recognition, surpassing even human-level capability. 3D object recognition errors rates are still high, but will rapidly evolve, especially with the use sensor fusion, e.g., ranging data from lidars. The camera-based technology will offer a more complete long-term roadmap toward improving the navigation. Many still utilize 2D lidar as it is easier to implement and is often good enough in known, controlled and slow-changing environments.

The cost of these forklifts is naturally higher, but the claimed ROI by many suppliers is within 12-18 months. The cost includes the installation and maintenance cost as well as the cost of the autonomous sensor suite, traction control and drivers, and the software, which can be amortized over a growing deployed fleet. Overall, price parity on an annual operational cost basis is nearly at hand in some high wage territories.

Over the past two years, this market has also entered into the early stages of its growth phase. Report authors believe this trend is likely to continue and accelerate, with the inflection point likely to be reached around 2025-2027. After this point, sales are projected to grow, exceeding 100k units by 2030. 

comments powered by Disqus