Chinese industrial robotics company Siasun has developed a new system for packing cigarettes which can cope with the ever-changing shapes and dimensions of tobacco products and packaging.
The company is using a delta robot to pack cigarettes into packaging which is increasingly unusual in the way it is designed, with some opening in the middle and other opening on the narrow side instead of the wide edge, which is the usual method globally. (See illustration below.)
The tobacco products themselves – mainly cigarettes – are also of different shapes and sizes.
There are two “standard” packaging boxes: 88mm long, 56mm wide and 23mm thick (hard); and 85mm long, 53mm wide and 22mm thick (soft).
The “standard” cigarette is 8 mm in diameter and 84 mm in length, and there are usually 20 cigarettes in each pack.
But beyond the standard dimensions and shapes for packaging and cigarettes, there are many new variations.
China has a population of approximately 1.6 billion people, and at least 350 million of them smoke, according to China Daily.
Developments within the tobacco industry and changes in consumer demand are leading to tobacco products of different shapes and dimensions, which are becoming more popular with consumers.
However, quickly and efficiently sorting and packing cigarettes of unusual shapes and dimensions has become a “major problem”, says Siasun, adding that the industry is “looking for to an efficient and intelligent sorting device that can completely solve this problem”.
Typically, a city-level tobacco logistics center has an average daily sorting volume of 500,000 to 600,000 pieces, a maximum peak sorting quantity of about 3 million pieces, and an average daily sorting quantity of shaped tobacco is 50,000 to 60,000 pieces, and the maximum sorting quantity is 200,000 pieces.
There are thousands of standard tobacco types, more than 300 types of special-shaped tobaccos with a high frequency of occurrence, and the largest type of about 600 types.
Siasun analysed the “pain points” or problems in the sorting of tobacco products. They found the following issues.
Shapes and sizes: There are many kinds of cigarettes with different shapes and sizes. Due to long-distance transportation, there is uncertainty in the angle, ordering and orientation of the cigarette case.
Identification: The height and size of the shaped smoke vary greatly, and the plastic film has serious reflection, making it difficult for the light source to take into account 300 to 600 types of smoke;
Palleting: The special-shaped cigarettes required by each order are different, so there is no effective algorithm for traditional type-shaped cigarettes to achieve the problem of stacking of special-shaped cigarettes;
Informatization: The confirmation and tracking of correct order information cannot be achieved during the transmission of traditional shaped cigarettes, and there are a large number of sorting errors that cause a large number of orders to be returned to the factory for unpacking.
After analysing the pain points, Siasun developed the “Xinsong Parallel Robot Intelligent Sorting Smoke Solution”.
With the Xinsong Delta parallel robot as the core, the system is equipped with Simsun visual inspection system Svision2D, which integrates two-dimensional visual inspection technology, intelligent fill light technology, deep learning algorithm technology, data fusion optimization technology, and provides a sorting and packaging system for shaped cigarettes.
Siasun says this is a “comprehensive and intelligent overall solution”, adding that the system implements the following functions:
- Open the data meridian to realize failure early warning: The Xinsong parallel robot sorting system can judge the status of incoming smoke through the visual system, adjust the order of smoking and the angle of the smoke, and combine the WMS order information to avoid order errors or damage to abnormal cigarettes caused by the disorder of incoming smoke. When an abnormality occurs, an alarm can be issued.
- Breakthrough of fill light technology to realize multi-species intelligent identification: The new vision system Svision2D intelligent fill light technology is inspired by the concept of shadowless lights in the operating room, which solves the situation that the robot body blocks the light source.
- Fusion of deep learning algorithms to optimize visual teaching: The new pine special-shaped cigarette sorting system incorporates deep learning algorithms, intelligently classifies cigarette recognition methods, dynamically adjusts recognition parameters, and optimizes the recognition success rate and recognition speed.
- Optimize data flow and improve system efficiency: This robot control system acquires order tasks in real time, optimizes the number of matching visual templates, and controls the recognition process to less than 100 milliseconds.
- Optimizing the special-shaped stacking algorithm to achieve dynamic special-shaped stacking: Automatically optimize the stacking algorithm according to the order information, and realize the automatic correction of the stacking shape when the order of the incoming smoke is chaotic.
Xinsong provides eight sets of special-shaped tobacco sorting and packaging parallel robot systems for tobacco logistics centers, which have all been put into use and processed tens of thousands of orders every day.
The person in charge of the technology of the logistics center said: “The original special-shaped tobacco sorting equipment had low efficiency and high error rates.
“Frequent errors, shutdowns, or incorrect orders were required, and workers needed to sort and pack online.
“During the period, they also looked for relevant companies and Experiments by research institutes were not satisfactory, and all ended in failure.”
Xinsong’s parallel robot special-shaped smoke sorting system “only takes short time”, says Siasun, and is “highly efficient” from commissioning to formal commissioning.
Siasun says it has “truly realized unmanned and automated sorting and packaging” for the cigarette industry.