COVID-19 is popular around the world. When dealing with COVID-19, the medical and health systems in every country are facing great challenges. No doubt, this round of COVID-19 will affect the development direction of the global health care system. At present, architects and medical and health consultants have discussed how to change the design of modern hospitals, so that the medical and health system can better cope with the crisis.
On the one hand, in the early days of the outbreak of COVID-19, a large number of medical staff were exposed to infected patients. In the epidemic situation, automatic equipment such as interrogation robot, disinfection robot and logistics robot have become just needed products of hospitals. It can be seen that for hospitals, in order to reduce the risk of infection in the daily work of hospital staff in the future, the public health system will develop in the direction of informatization and automation.
Based on the vision of reducing the risk of infection of nurses in infection department, bOnce robot from Harbin Institute of technology intelligent central research institute has developed a medical puncture blood collection robot under the guidance of ultrasound, which realizes the automatic collection and sorting of blood samples, and helps the hospital to realize full automatic inspection.
Hagong intelligent is a high-tech listed company focusing on high-end intelligent equipment manufacturing and artificial intelligent robots in China. It has profound experience in robot research and development, and takes AI + robot as the core development concept to improve the robot life circle.
Although Harbin Industrial intelligence has completed the layout of the whole industrial chain in terms of industrial robots, medical and industrial industries are very different. How can boce robot transplant the experience of industrial robot research and development to the medical field? How does the blood collection robot really land in the medical scene? Arterial network interviewed Xie Lei, CEO of boce robot, and Zhang Zhaodong, CTO.
What kind of pain point does the most common blood sampling have
In the past, bOnce robot team has been dealing with robots for a long time, but in their view, the development of the company must attach equal importance to social responsibility. The accumulated technical direction should be used in the most needed scenarios.
“We finally chose the field of medical robots as a new field of research and development, and used our accumulated experience in the field of robots to solve the problems in the medical field.”
In the investigation, the team of bOnce robot found that the application of medical robot can be roughly divided into four directions: first, robot + treatment direction; The second is robot + diagnosis; The third is the direction of monitoring; The fourth is the direction of inspection. Xie Lei believes that in the field of treatment, many head enterprises have gathered, and intuitive surgery, which belongs to Da Vinci robot, is the best. In the field of diagnosis, AI software is more preferred. The field of monitoring is mainly hardware equipment. Finally, diagnosis and detection, which is a relatively blank field of intelligent robot application, is chosen as the entry point.
At first, the robot team found that the most easy link for the laboratory to combine with the robot is the pretreatment of the blood test tube after blood collection. The shaking and labeling of the test tube can quickly transplant the industrial robot products to the medical scene. But what the boce robot team wants to do is products that are closely integrated with the medical scene, rather than simply copying products from industry to the medical field.
In the investigation, the robot team further found that in the blood collection process, nurses are at great risk of infection due to needle pricking. “We want to develop blood collection robots to reduce the risk of infection among nurses, especially those in infection departments and infectious disease hospitals,” Xie said. We have conducted a survey on medical staff and found that the annual average number of needle stick injuries per capita is 0.75, and the work and prevalence rate of doctors under radiation exceeds 60%. “
In addition to the risk of accidents, the workload of nurses blood collection is also very large. Xie Lei, CEO of boce robot, told arteria.com: “in the top 100 Grade-A hospitals, the average daily blood collection times reached 6000 people; The average daily service of a single medical staff is 150 person times, and the laboratory technician has to act as a blood sampling nurse in many hospitals. “
Three dimensional imaging ensures puncture accuracy
Needling is a required course for nurses. But for robots, if they want to learn and complete the needling action, they need to have human “visual ability” and “judgment ability”.
Zhang Zhaodong said that the medical scene has high requirements for robots, and the team of bOnce has a deep accumulation in the fields of mechanical control and artificial intelligence. The main difficulty that the team of bOnce overcomes is whether it can achieve accurate puncture.
“Because we have a deep accumulation in the field of robotics, at the beginning, we thought that the design of the blood collection robot was not difficult. We just need to simulate the process of nurses’ needling and blood collection, and use the mechanical control principle to complete it. After the real research and design, we found that the real problem lies in the selection of blood vessels and puncture control. Because the degree of stenosis, depth and direction of blood vessels are different, the depth of puncture is different. In the design of industrial robots, we only need to make robots from point a to point B, but in the design of blood collection robots, we need robots to make accurate judgments and decisions. “
In order to make the blood collection robot be able to see through the blood vessels, bOnce robot has studied many machine vision AI technologies, integrating the two technologies to achieve accurate puncture.
First, infrared imaging technology, through the reflection of hemoglobin in blood vessels to infrared light, collects the data of blood vessel direction, and then applies ultrasonic imaging.
“The infrared imaging image is like a plane GPS. Although we can see the specific coordinates in the sky, we can’t judge whether there is something underground. In the process of blood collection, nurses can judge the depth of the needle into the blood vessel by feeling, but for the machine, it is not perceptible. “
In order to enable the machine to perceive the existence of blood vessels stereoscopically and accurately, the robot decided to add ultrasound images to the imaging module.
In addition to the ultrasonic images, this makes the robot team of boce encounter a lot of difficulties. Most of the whole technical team have no experience in analyzing medical ultrasonic images. Through many medical ultrasound experts in the industry, we have added rich relevant experience and achieved long-term cooperation in puncture application under medical ultrasound images.
“We all know that ultrasound requires medical staff to hit in the right place to get the image, not that ultrasound can get the image anywhere. In the blood collection robot, the ultrasonic probe must have a certain angle with the blood vessel to obtain the image. We put a lot of effort into translating these requirements into engineering language. “
The application of ultrasound image is the most time-consuming part of the R & D investment of the brce team. With the ultrasonic image, we can grasp the image of the vascular section, obtain the information of the diameter and depth of the blood vessel, and help the blood collection robot to establish the three-dimensional coordinates of the blood vessel.
Combined with infrared imaging and ultrasonic imaging, the robot can realize fast 3D image positioning under the assistance of ultrasonic positioning, and quickly calculate the optimal blood vessels according to the imaging blood vessel distribution map. In the aspect of vascular segmentation, the diameter and position of the artery and vein are detected and segmented by ultrasound image.
In the space motion control, through the fusion of image and ultrasound, the robot can calculate the trajectory and complete the puncture and blood collection.
The superposition of the two image recognition technologies greatly improves the accuracy of the robot puncture. Through accurate calculation, the blood sampling robot can control the needle entry angle and puncture depth of blood sampling puncture, and bring less pain to patients than manual puncture.
In order to make the simulation degree of blood sampling robot puncture and nurse blood sampling higher, the puncture force feedback control module is added to the robot, which can real-time monitor the puncture force distance, the data change of motor stall, and assist the closed-loop algorithm of vision and moving parts, so as to better improve the stability of puncture.
“In the area of force feedback, we also spent a lot of time in research and development. We asked the nurses how to get to the point. The nurses told us that human blood vessels have a special feeling under the skin. When the needle enters the blood vessels, there will be a” sense of emptiness. “. It’s hard for machines to capture this particular feeling. We’ve come up with a way to add a lot of sensors to predict changes in power. “
At present, the product of boce robot is still in the process of research and development iteration, and now the product of boce can complete blood collection in one minute. At present, the blood collection robot of boce can be divided into six systems
Puncture blood collection system: it can guide puncture after AI calculation of medical image;
Electronic pulse pressing system: it can replace the contact film and compress the arm with constant force;
Test tube management system: realize test tube labeling, test tube sorting, test tube shaking, etc;
Consumables management system: automatic replacement of needles, disinfectants and couplants;
Hemostatic nursing system: automatic hemostatic compression treatment;
Human computer interaction system: through his / LIS system, realize automatic data check, at the same time, demonstration teaching display.
Through the six systems covering all aspects of the blood collection process. Ensure the standardization of puncture and blood collection, ensure that the whole process meets the blood collection standard and monitor the error proofing link; Avoid cross infection and needle stick injury of patients and medical staff. In terms of automation, the whole process of blood collection and puncture is independent, and no intervention is needed. The superposition of a variety of technologies also ensures that the products of boce robot can deal with obese patients, elderly patients and other special physical patients.
Hope to first land in infectious disease hospital
On the landing scene, Xie Lei said that as a new product, the boce robot team hopes that the earliest landing will be in infectious disease hospitals and departments.
“The outbreak of covid-19 in the world shows us the horror of infectious diseases. Fever clinic is a very high-risk work scene. If we manage fever clinic according to strict isolation and infection reduction standards. So the blood collection robot can reduce a lot of risk. Relying on the intelligence of Harbin Institute of technology, the primary motivation for us to make this product is to consider the social value. “
Nurses often need to complete blood collection in multi scene mobile situation. Therefore, the portable blood collection robot is also being developed. In the world, blood collection is one of the most common medical procedures, blood collection is very common does not mean it is a simple thing.
Editor in charge: GT