Large-scale penetration of mobile robots requires the "human-machine integration" test

"Young people have a high turnover rate, high labor costs, reduced efficiency, difficult management, and high investment. The manufacturing market is highly competitive, and replacing people with machines is the inevitable trend of industrial logistics. Faced with a stock of over one million vehicles and a market of hundreds of billions, light and small industrial vehicles will be the main force of industrial logistics. "

Yao Haijin, founder and CEO of Dolphin Star, said: "For mobile robot companies, if they want to achieve large-scale industry applications, they must develop their products in the direction of intelligence, and simple use, human-machine collaboration, and cluster operations are the inevitable evolution direction. "

Dolphin Star was founded in 2017 and is committed to the research and development and manufacturing of light and small industrial mobile robot products. It has established a complete intelligent factory robot handling system. Innovative AGV unmanned forklift modular structure design, mini laser-guided unmanned transport vehicle, laser plus visual hybrid guidance technology , etc. It has independent core technologies in SLAM laser navigation technology, motion control, visual recognition control, and remote dispatching system.

On August 27, Dolphin Star held the 2022 AiTEN Robot Scene Demonstration and Channel Partner Meeting in Suzhou. Faced with the difficulties of human-machine integration in the dynamic and complex scenes of industrial logistics, Dolphin Star gave its own answer.

Human-machine perception

In the early days, robots needed to be programmed before they could perform their intended actions. The future trend of robots is to make self-judgments as the environment changes, without the need for prior intervention .

In this case, the opportunities for robots to interact with the environment increase, and the technical requirements for sensors also increase. Currently, the more commonly used sensors are two-dimensional laser sensors and visual sensors.

Two-dimensional laser sensors are very stable, can measure long distances and a large range, and have high measurement accuracy. They are a relatively mature solution so far. Their disadvantage is that they are too expensive.

Visual sensors can perceive a wide range of environmental information and are not expensive. Currently, there are many intelligent control algorithms based on visual sensors in the industry. However, visual sensors are not very mature because the visual information is too rich and the environment changes too much.

In response to this dilemma, Dolphin Star has developed laser plus visual hybrid guidance technology and designed a high-performance perception controller . It can achieve efficient positioning and mapping in complex scenes, realize dynamic visual human target recognition and intelligent target following, adapt to complex scenes, and perform real-time dynamic planning for changing scenes.

At present, the industrial mobile robot products provided by Dolphin Star have covered the three major scenarios of unmanned handling, warehousing and distribution within the factory.

Human-machine decision making

The construction of unmanned factories is not achieved overnight. Especially in the current industrial scenario, experienced forklift operators are much more efficient than handling robots.

Dolphin Star's intelligent management and scheduling system uses cloud computing technology to quickly generate node task allocation plans and perform warehouse status maintenance, thereby assisting managers in decision-making and achieving on-site safety maintenance .

The intelligent management and dispatching system has five major characteristics:

1. ε-greedy decision evaluation . Using the experience replay pool, evaluate the network reinforcement learning performance.

2. Incremental ELM traffic control prediction model . It can autonomously identify the learning environment and support multiple operation modes with the optimal traffic control strategy.

3. Predictive scheduling of AGV resources and traffic resources . Predict supply and demand differences and coordinate cluster operations of more than 100 units with the best real-time solution.

4. Distributed cloud scheduling architecture . Single-machine task self-control + cloud-based task aggregation and multi-level linkage of operation scenarios.

5. Open interface design . Standardized communication protocols and interfaces support mixed scheduling of various AGVs and field equipment.

Today, all Dolphin Star series products can be equipped with intelligent management and scheduling systems to meet the needs of 3C electronics, warehousing, manufacturing, auto parts, food and medicine and other fields.

Human-machine operation

At present, the unmanned forklift market has basically achieved customer cultivation and testing, and is waiting for the emergence of low-priced and stable products.

Dolphin Star has created a modular forklift product matrix including M series handling robots, T series distribution robots, and A series storage robots . It can achieve rapid scenario customization, with a large parameter performance range, high reliability, low debugging cost, and short deployment cycle, helping terminal companies improve operating efficiency and reduce operating costs.

In 2019, Dolphin Star launched a product suitable for various material forms (pallets, material frames, material carts, etc.), which is often used in industrial manufacturing process material distribution, transit material warehouses and other places where frequent transportation of goods is required - the "Little Dolphin" handling robot MP10S.

"Little Dolphin" is a mini AGV designed for places such as smart factories that require frequent internal transportation of goods. It is light, flexible and versatile. It is 1475mm long, 735mm wide and 450mm high, with a turning radius of only 1m and a rated load of up to 1T. In March this year, "Little Dolphin" won the 2022 German Red Dot Product Design Award.

Compared with traditional unmanned forklifts, the "Little Dolphin" has the driving mechanism designed into the fork legs. It has a compact structure and is mini and small, which is particularly suitable for narrow channel scenarios. It can pass through channels that were previously only accessible to jack-up AGVs, thereby greatly improving the utilization rate of storage space.

This product is also Dolphin Star's universal, lightweight and popular product. It is a tireless "smart porter" of the new era . It can cover the application needs of all scenarios with standard products. It has now entered the large-scale shipment period.

Human-machine safety

To achieve human-machine integration, safety is a key issue that cannot be ignored. Dolphin Star uses many safety devices to prevent accidents.

First, the dispatching system traffic control . Dolphin Star uses the dispatching system to plan in advance. The traffic control in its system will prevent AGVs from running into each other, and there is a concept of traffic lights, which ensures safety and smooth passage between AGVs and people inside the factory.

The second is obstacle avoidance laser . Each Dolphin Star AGV is equipped with a 180° obstacle avoidance laser. When an obstacle within 1m is identified, the vehicle will slow down; when an obstacle within 0.5m is identified, the vehicle will stop.

The third is physical anti-collision . The Dolphin Star will be equipped with an emergency stop button on the body and an anti-collision strip at the front to provide layered safety protection.

Fourth, visual equipment . The Dolphin Star AGV body can also be equipped with visual equipment, and visual navigation can maximize safety.

In Yao Haijin's view, AiTEN, pronounced like "爱折腾", means Dolphin Star is a company that "loves tossing". "We are currently just accumulating products and talents. Next, Dolphin Star will challenge and toss in higher technical fields to enable more factories to realize digital intelligent manufacturing."