Deliver in half an hour
In previous articles, we talked about the development of Sberbank's direction of robotics , collaborative robots and exoskeleton . In this talk about logistic robots.
Photo: Sberbank
The International Federation of Robotics classifies logistics robots to the class of professional service robots and calls them logistic systems.
According to ISO 8373: 2012 , a professional service robot is a robot that performs useful work for people and equipment, excluding industrial automation tasks, and is used to reap the benefits of various services.
Thus, the logistic robot is a mobile programmable drive mechanism designed to control the flow of goods, their transportation, processing and packaging. As a rule, logistic robots are installed in:
Logistic robots have been developed due to the fact that they can reduce costs and improve the efficiency of enterprises.
In an industrial environment, robots can reduce the need for human labor, reduce the number of human errors, improve safety at workplaces, and improve the accuracy of inventory and accounting of goods / parts by eliminating a person from performing routine tasks and automating logistics processes.
Robots for industrial environments can be used in flexible industries / enterprises to solve a wide range of tasks:
Courier-type robots are used in non-production environments that benefit both business owners and their customers thanks to:
The robot courier usually performs the following actions:
The B2C segment accounts for about 50% of the total market for the delivery of goods, according to a McKinsey & Company survey . It is assumed that the use of unmanned delivery vehicles will reduce the cost of B2C delivery by half, according to the same study. For customers of e-commerce services, delivery of goods by robots will bring an interesting feature - the ability to choose the exact time of delivery of goods through the store application (if there is such a function and free robots nearby).
In the future, unmanned vehicles for the transportation of goods will reduce the cost of wages to drivers, increase the efficiency of cargo delivery due to the possibility of drones to continue to move without stopping for rest, increase safety of transport by eliminating the human factor.
Unmanned delivery vehicles can be used throughout the supply chain for day-to-day logistics: from the “warehouse of the future” to unmanned vehicles and unmanned aerial vehicles to overcome the problem of the last mile. UAVs can carry out:
Each year, the bank processes about eighty thousand tons of banknotes: banknotes are packaged in bundles, packs are packaged in cassettes, cartridges are moved from one point to another, and all movements are performed by people. By automating at least part of the process, you can significantly reduce the cost of these operations, increase productivity, account accuracy and reduce the number of errors.
Another vivid example is the delivery of correspondence and cargo for bank employees and customers. Clients need to deliver the cards and take the paper copies of the documents signed by them to the storage centers. For each employee, the logistics service of the bank delivers 20 kg of cargo per month. Taking into account the geographical features of our country and the huge number of Sberbank branches throughout the country, the task of accurate and uninterrupted cargo delivery becomes very interesting for robotization.
That is why logistics robots are one of the priorities of the Sberbank robotics laboratory. The other day, we closed the set to the corporate accelerator (applications were accepted, including logistic systems). Be sure to tell about interesting projects.
The market of logistics robots cannot boast of the volume of the industrial manipulator market: in 2016, the number of installed logistics systems increased by 34% compared to 2015, but amounted to only 25,444 units with a total value of about $ 0.9 billion (against 1.8%). million units of industrial robots with a market volume of about $ 13 billion).
It is predicted that 190,000 logistics robots will be sold between 2018 and 2020, with an average annual market growth of between 25% and 30%. At the same time, logistics robots will account for 48% of sales of all service robots in 2018-2020.
Figure - Sales of logistic robots in 2015-2016, as well as the 2017 forecast for 2017 and 2018-2020 ( source ) The
growth drivers of the logistics logistics market, the development of artificial intelligence technologies, new materials, as well as cost reduction of robots by reducing costs parts:
Logistic robots began their march relatively recently and much later than industrial counterparts. Until recently, only industrial robots were widespread. They could not move, were blind, relatively unreasonable, and performed the same movements over and over thousands of times a day with a high degree of accuracy and repeatability. For many simple manufacturing processes, such as welding or transferring parts, these skills are necessary. However, to solve logistical tasks of high accuracy and repeatability is not enough: the robot must not only act, but also understand the world around it, build behavior patterns depending on dynamically changing conditions, make decisions in the most optimal way.
The logistic robot has to handle a wide range of different products / parts in an infinite number of combinations. Despite the complexity of the task, it is achievable if the robot can possess the abilities mentioned above: to feel the world, build models, make decisions and act according to circumstances.
The first experimental samples of intelligent mobile robots began to appear around the 1960s.
One of the first was the robot Shakey the Robot , developed at the Center for Artificial Intelligence at the Stanford Research Institute (now called SRI International).
Figure - Shakey Robot ( source ) The
robot was equipped with a television camera, an optical rangefinder, and collision sensors. He knew how to move in office spaces and to distinguish specially colored objects.
Shakey the robot was the first general-purpose mobile robot with the beginnings of artificial intelligence. While other robots need to be instructed at every single stage of a larger task, Shakey could analyze the commands and break them into main parts on their own, thanks to a specially designed planner .
The Shakey project combined research in the fields of robotics, computer vision and natural language processing. The most notable project results areA * search algorithm , Hough transform and visibility graph method .
Another interesting representative of early mobile robots with artificial intelligence is the stand-alone mobile robot Hilare. The Hilare robot was developed in 1977 at the Systems Analysis and Architecture Lab in Toulouse, France. He is considered to be the first French mobile robot capable of working autonomously in an unknown environment, perceiving and analyzing the world around him and making decisions based on his analysis.
Figure - Hilare Robot ( source )
Hilare already looks more like a modern mobile / logistics robot. So, Hilare was equipped with:
Subsequent versions of this research robot, Hillare 2 (1990) and Hillare 3 (1999), led to significant progress in the field of mobile robots and brought the appearance of modern logistic devices closer.
Picture - Hilare 2 Robot ( source )
One of the first courier robots that became widespread in hospitals was TRC's Helpmate , developed in the early 1990s. In total, about 100 robots were commissioned. Helpmate delivered food, medicine, etc. to patients. The main success of the project was the release of qualified personnel from the implementation of courier tasks. The system was ahead of its time and successfully paved the way for several courier systems in hospitals and indoors.
Figure - TRC Helpmate ( source )
Despite the high number of suppliers of warehouse robots, fully autonomous warehouse robots (able to move without special preparation of the premises) came into use relatively recently. Many companies are actively developing in this area.
For example, OmniRob from KUKA (Germany) implements the concept of a mobile manipulator for solving problems of capturing and carrying parts and goods. The robot is a platform with a manipulator with a gripping device mounted on it. A distinctive feature of the platform are its wheels - the so-called Omni-wheels , which have the ability to move in different directions.
Figure - KUKA OmniRob ( source )
Another mobile manipulator MM-800 from Neobotix (Germany) has a KUKA six-step manipulator mounted on a mobile platform. The platform provides autonomous operation for 10 hours. MM-800 is used at the AUDI plant in Ingolstadt.
Figure - Neobotix MM-800 ( source )
A striking example of the successful creation and implementation of warehouse logistics robots can serve as the company Kiva Systems , later bought by Amazon and renamed Amazon Robotics. The company was founded in 2003 and already by 2008 introduced its solution to three major customers.
Figure - Amazon (Kiva) AGV ( source ) The
company's robots are mobile platforms with a carrying capacity of 500 and 1500 kilograms, equipped with infrared sensors to avoid collisions, an on-board computer and wireless communication with cloud management software. All robots have batteries on board and must be recharged once an hour for five minutes.
To meet the challenges of speeding up and simplifying the delivery of goods, Kiva has developed its own approach to organizing a warehouse. All products are stored in specialized racks. When you enter an order, the system finds the nearest robot and sends it to the rack with the necessary goods. Robots move through the warehouse following QR codes placed on the floor. When the robot reaches a given point, it enters under the rack, lifts it with the help of a special screw mechanism and carries it to the operator to deliver the goods.
The SwissLog SpeciMinder and RoboCourier robots are compact stand-alone mobile platforms that move freely in typical stationary and laboratory environments. Platforms are equipped with places for installation of goods and can autonomously deliver goods to the specified recipients.
Drawing - SwissLog RoboCourier ( source )
Robots are equipped with interfaces for integration with doors and elevators to move between rooms and floors, a laser navigation system for dynamically re-planning paths and avoiding obstacles. SpeciMinder and RoboCourier can work autonomously - they do not need a controlling server, navigation and planning tasks can be solved on board the robots.
The company Aethon (USA), offers a different approach. It is based on a compact mobile platform to which carts can be docked for storing and transporting various hospital goods (plates, medicines, bedding, blood samples, pharmaceutical preparations, etc.). Platforms can dock to all supported trucks to reduce traffic and increase system efficiency (for example, a trolley can be transported in a shuttle mode: one platform brings the trolley to a certain point, then another platform carries the cart). It is alleged that in operation there are more than 500 robots in about 450 hospitals that perform more than 5 million deliveries per year.
Figure - Aethon TUG ( source )
Numerous startups develop and offer conceptually similar ground or flying unmanned aerial vehicles to reduce the cost and speed up the delivery of goods to the end user. Among them: SideWalk (partnership with DHL), Starship (created by founding engineers of Skype) and Dispatch (MIT / UPenn).
Figures - Starship and Dispatch Carry Robots
A starship robot can carry a load of up to 10 kg, while a Dispatch robot can carry loads of up to 50 kg. These robots are used to deliver goods no further than 30 minutes from the local distribution center, and are capable of delivering small packages, groceries, pizza, etc.
Among the development of delivery systems using unmanned aerial vehicles can be identified Amazon Prime Air Delivery , Drone Delivery Canada and JD.COM . All of them, with varying degrees of success, work out the same case - the quick delivery of small-sized goods and weight from online stores to customers using unmanned aerial vehicles. Amazon made the first commercial delivery on December 17, 2016. Drone Delivery Canada is working out its drone management and dispatching system at one of the airports in the United States, JD.COM began to deliver goods to rural regions as part of trial operation.
Figure - Amazon Prime Air , Canada Drone Delivery , JD.COM
Where to look and in what direction to develop your own product?
Robots live in the digital world, and for their more efficient and meaningful work, it is necessary to maximally extend the digital world to the physical. New concepts of digitizing media and networking them (IoT, Industry 4.0) can raise enterprise automation to a significantly higher level.
Improving performance, speed and fully autonomous navigation without special preparation of the premises will help reduce the threshold for the product to reach customers.
Increased availability of the system by reducing the charging time of the robots, as well as improving the reliability of the devices will help to attract customers.
Quick and intuitive setup and configuration, ease of maintenance, ease of use will help reduce the threshold of entry and attract customers.
Robots have already learned to drive independently, but they still have problems with pattern recognition in difficult conditions. The development of perceptual technologies will allow robots to produce a full cycle of assembling orders from warehouses without human intervention. This will reduce the time of preparation of orders and the cost of maintaining the warehouse.
The ability to work alongside staff and perform tasks in the same space with a person allows you to customize production and logistic processes.
Photo: Sberbank
What is a logistic robot?
The International Federation of Robotics classifies logistics robots to the class of professional service robots and calls them logistic systems.
According to ISO 8373: 2012 , a professional service robot is a robot that performs useful work for people and equipment, excluding industrial automation tasks, and is used to reap the benefits of various services.
Thus, the logistic robot is a mobile programmable drive mechanism designed to control the flow of goods, their transportation, processing and packaging. As a rule, logistic robots are installed in:
- industrial environments for moving objects of all kinds between machine tools, transfer points or warehouses;
- non-production environments, such as warehouses, postal and logistics centers, in hospitals or other public buildings for the transportation, delivery and transfer of goods;
- in open areas: in ports, airports and transshipment centers, as well as to overcome the “last mile” problem in delivering goods to end users
What is their value, what is useful?
Logistic robots have been developed due to the fact that they can reduce costs and improve the efficiency of enterprises.
In an industrial environment, robots can reduce the need for human labor, reduce the number of human errors, improve safety at workplaces, and improve the accuracy of inventory and accounting of goods / parts by eliminating a person from performing routine tasks and automating logistics processes.
Robots for industrial environments can be used in flexible industries / enterprises to solve a wide range of tasks:
- moving products through production processes, ensuring delivery just-in-time (JIT): delivery of parts to the processing / assembly point at the right time;
- delivery of parts for assembly;
- order picking: moving ordered products to the trailer loading area for shipment;
- just-in-time delivery (JIT): delivery of parts to the processing / assembly point at the right moment;
Courier-type robots are used in non-production environments that benefit both business owners and their customers thanks to:
- reduction of traffic in warehouses;
- reducing the risk of accidents and congestion;
- increase the reliability of the enterprise as a whole and eliminate errors related to human factors;
- increase enterprise productivity;
- increased availability and service flexibility, reduced transportation costs and time (for customers).
The robot courier usually performs the following actions:
- detection / detection of objects for transportation (sometimes with the help of a person by pressing the robot call button);
- loading and transportation of objects;
- optimization and adaptation of transport routes;
- classification and accounting of objects (allows you to improve the accuracy of accounting of material resources);
- unloading of objects in the destination zone;
The B2C segment accounts for about 50% of the total market for the delivery of goods, according to a McKinsey & Company survey . It is assumed that the use of unmanned delivery vehicles will reduce the cost of B2C delivery by half, according to the same study. For customers of e-commerce services, delivery of goods by robots will bring an interesting feature - the ability to choose the exact time of delivery of goods through the store application (if there is such a function and free robots nearby).
In the future, unmanned vehicles for the transportation of goods will reduce the cost of wages to drivers, increase the efficiency of cargo delivery due to the possibility of drones to continue to move without stopping for rest, increase safety of transport by eliminating the human factor.
Unmanned delivery vehicles can be used throughout the supply chain for day-to-day logistics: from the “warehouse of the future” to unmanned vehicles and unmanned aerial vehicles to overcome the problem of the last mile. UAVs can carry out:
- logistics in the sorting centers: ports, airports, railway hubs, customs centers;
- delivery of goods from production sites to distribution centers;
- delivery of goods to final consumers.
Take, for example, Sberbank
Each year, the bank processes about eighty thousand tons of banknotes: banknotes are packaged in bundles, packs are packaged in cassettes, cartridges are moved from one point to another, and all movements are performed by people. By automating at least part of the process, you can significantly reduce the cost of these operations, increase productivity, account accuracy and reduce the number of errors.
Another vivid example is the delivery of correspondence and cargo for bank employees and customers. Clients need to deliver the cards and take the paper copies of the documents signed by them to the storage centers. For each employee, the logistics service of the bank delivers 20 kg of cargo per month. Taking into account the geographical features of our country and the huge number of Sberbank branches throughout the country, the task of accurate and uninterrupted cargo delivery becomes very interesting for robotization.
That is why logistics robots are one of the priorities of the Sberbank robotics laboratory. The other day, we closed the set to the corporate accelerator (applications were accepted, including logistic systems). Be sure to tell about interesting projects.
Market condition
The market of logistics robots cannot boast of the volume of the industrial manipulator market: in 2016, the number of installed logistics systems increased by 34% compared to 2015, but amounted to only 25,444 units with a total value of about $ 0.9 billion (against 1.8%). million units of industrial robots with a market volume of about $ 13 billion).
It is predicted that 190,000 logistics robots will be sold between 2018 and 2020, with an average annual market growth of between 25% and 30%. At the same time, logistics robots will account for 48% of sales of all service robots in 2018-2020.
Figure - Sales of logistic robots in 2015-2016, as well as the 2017 forecast for 2017 and 2018-2020 ( source ) The
growth drivers of the logistics logistics market, the development of artificial intelligence technologies, new materials, as well as cost reduction of robots by reducing costs parts:
- Artificial intelligence allows to lower the entry threshold for robots: they no longer need to prepare premises for them in a special way, navigation can be carried out without special markers.
- New materials and the development of energy storage technologies reduce the downtime of robots, increasing their attractiveness.
- Mass production of components allows manufacturers to reduce prices, which has a positive effect on the cost of the entire solution as a whole.
The history of development
Logistic robots began their march relatively recently and much later than industrial counterparts. Until recently, only industrial robots were widespread. They could not move, were blind, relatively unreasonable, and performed the same movements over and over thousands of times a day with a high degree of accuracy and repeatability. For many simple manufacturing processes, such as welding or transferring parts, these skills are necessary. However, to solve logistical tasks of high accuracy and repeatability is not enough: the robot must not only act, but also understand the world around it, build behavior patterns depending on dynamically changing conditions, make decisions in the most optimal way.
The logistic robot has to handle a wide range of different products / parts in an infinite number of combinations. Despite the complexity of the task, it is achievable if the robot can possess the abilities mentioned above: to feel the world, build models, make decisions and act according to circumstances.
The first experimental samples of intelligent mobile robots began to appear around the 1960s.
Shakey
One of the first was the robot Shakey the Robot , developed at the Center for Artificial Intelligence at the Stanford Research Institute (now called SRI International).
Figure - Shakey Robot ( source ) The
robot was equipped with a television camera, an optical rangefinder, and collision sensors. He knew how to move in office spaces and to distinguish specially colored objects.
Shakey the robot was the first general-purpose mobile robot with the beginnings of artificial intelligence. While other robots need to be instructed at every single stage of a larger task, Shakey could analyze the commands and break them into main parts on their own, thanks to a specially designed planner .
The Shakey project combined research in the fields of robotics, computer vision and natural language processing. The most notable project results areA * search algorithm , Hough transform and visibility graph method .
Hilare
Another interesting representative of early mobile robots with artificial intelligence is the stand-alone mobile robot Hilare. The Hilare robot was developed in 1977 at the Systems Analysis and Architecture Lab in Toulouse, France. He is considered to be the first French mobile robot capable of working autonomously in an unknown environment, perceiving and analyzing the world around him and making decisions based on his analysis.
Figure - Hilare Robot ( source )
Hilare already looks more like a modern mobile / logistics robot. So, Hilare was equipped with:
- ultrasonic sensors to detect nearby objects;
- laser rangefinder for making a two-dimensional map of the environment;
- odometer to measure the distance traveled;
- Four Intel 80286 microprocessors for processing data from sensors.
Subsequent versions of this research robot, Hillare 2 (1990) and Hillare 3 (1999), led to significant progress in the field of mobile robots and brought the appearance of modern logistic devices closer.
Picture - Hilare 2 Robot ( source )
TRC Helpmate
One of the first courier robots that became widespread in hospitals was TRC's Helpmate , developed in the early 1990s. In total, about 100 robots were commissioned. Helpmate delivered food, medicine, etc. to patients. The main success of the project was the release of qualified personnel from the implementation of courier tasks. The system was ahead of its time and successfully paved the way for several courier systems in hospitals and indoors.
Figure - TRC Helpmate ( source )
Examples of modern systems
Logistic robots for production environments
Despite the high number of suppliers of warehouse robots, fully autonomous warehouse robots (able to move without special preparation of the premises) came into use relatively recently. Many companies are actively developing in this area.
KUKA OmniRob
For example, OmniRob from KUKA (Germany) implements the concept of a mobile manipulator for solving problems of capturing and carrying parts and goods. The robot is a platform with a manipulator with a gripping device mounted on it. A distinctive feature of the platform are its wheels - the so-called Omni-wheels , which have the ability to move in different directions.
Figure - KUKA OmniRob ( source )
Neobotix MM-800
Another mobile manipulator MM-800 from Neobotix (Germany) has a KUKA six-step manipulator mounted on a mobile platform. The platform provides autonomous operation for 10 hours. MM-800 is used at the AUDI plant in Ingolstadt.
Figure - Neobotix MM-800 ( source )
Logistic robots for non-production environments
Amazon Robotics (Kiva systems)
A striking example of the successful creation and implementation of warehouse logistics robots can serve as the company Kiva Systems , later bought by Amazon and renamed Amazon Robotics. The company was founded in 2003 and already by 2008 introduced its solution to three major customers.
Figure - Amazon (Kiva) AGV ( source ) The
company's robots are mobile platforms with a carrying capacity of 500 and 1500 kilograms, equipped with infrared sensors to avoid collisions, an on-board computer and wireless communication with cloud management software. All robots have batteries on board and must be recharged once an hour for five minutes.
To meet the challenges of speeding up and simplifying the delivery of goods, Kiva has developed its own approach to organizing a warehouse. All products are stored in specialized racks. When you enter an order, the system finds the nearest robot and sends it to the rack with the necessary goods. Robots move through the warehouse following QR codes placed on the floor. When the robot reaches a given point, it enters under the rack, lifts it with the help of a special screw mechanism and carries it to the operator to deliver the goods.
SwissLog RoboCourier
The SwissLog SpeciMinder and RoboCourier robots are compact stand-alone mobile platforms that move freely in typical stationary and laboratory environments. Platforms are equipped with places for installation of goods and can autonomously deliver goods to the specified recipients.
Drawing - SwissLog RoboCourier ( source )
Robots are equipped with interfaces for integration with doors and elevators to move between rooms and floors, a laser navigation system for dynamically re-planning paths and avoiding obstacles. SpeciMinder and RoboCourier can work autonomously - they do not need a controlling server, navigation and planning tasks can be solved on board the robots.
Aethon tug
The company Aethon (USA), offers a different approach. It is based on a compact mobile platform to which carts can be docked for storing and transporting various hospital goods (plates, medicines, bedding, blood samples, pharmaceutical preparations, etc.). Platforms can dock to all supported trucks to reduce traffic and increase system efficiency (for example, a trolley can be transported in a shuttle mode: one platform brings the trolley to a certain point, then another platform carries the cart). It is alleged that in operation there are more than 500 robots in about 450 hospitals that perform more than 5 million deliveries per year.
Figure - Aethon TUG ( source )
Logistic robots for open areas
Numerous startups develop and offer conceptually similar ground or flying unmanned aerial vehicles to reduce the cost and speed up the delivery of goods to the end user. Among them: SideWalk (partnership with DHL), Starship (created by founding engineers of Skype) and Dispatch (MIT / UPenn).
Figures - Starship and Dispatch Carry Robots
A starship robot can carry a load of up to 10 kg, while a Dispatch robot can carry loads of up to 50 kg. These robots are used to deliver goods no further than 30 minutes from the local distribution center, and are capable of delivering small packages, groceries, pizza, etc.
Among the development of delivery systems using unmanned aerial vehicles can be identified Amazon Prime Air Delivery , Drone Delivery Canada and JD.COM . All of them, with varying degrees of success, work out the same case - the quick delivery of small-sized goods and weight from online stores to customers using unmanned aerial vehicles. Amazon made the first commercial delivery on December 17, 2016. Drone Delivery Canada is working out its drone management and dispatching system at one of the airports in the United States, JD.COM began to deliver goods to rural regions as part of trial operation.
Figure - Amazon Prime Air , Canada Drone Delivery , JD.COM
New technologies
Where to look and in what direction to develop your own product?
Iot
Robots live in the digital world, and for their more efficient and meaningful work, it is necessary to maximally extend the digital world to the physical. New concepts of digitizing media and networking them (IoT, Industry 4.0) can raise enterprise automation to a significantly higher level.
Autonomy
Improving performance, speed and fully autonomous navigation without special preparation of the premises will help reduce the threshold for the product to reach customers.
Availability
Increased availability of the system by reducing the charging time of the robots, as well as improving the reliability of the devices will help to attract customers.
Simplicity
Quick and intuitive setup and configuration, ease of maintenance, ease of use will help reduce the threshold of entry and attract customers.
Artificial Intelligence
Robots have already learned to drive independently, but they still have problems with pattern recognition in difficult conditions. The development of perceptual technologies will allow robots to produce a full cycle of assembling orders from warehouses without human intervention. This will reduce the time of preparation of orders and the cost of maintaining the warehouse.
Collaborative
The ability to work alongside staff and perform tasks in the same space with a person allows you to customize production and logistic processes.