UK-China Agritech Challenge: Envirobot An autonomous roving platform for environment, health and welfare monitoring of poultry

Lead Research Organisation: Royal Veterinary College
Department Name: Pathobiology and Population Sciences

Abstract

The poultry industry in China is modernising at a fast pace, but in the process is encountering significant problems with controlling temperature and other key parameters within their production facilities. The technologies currently used in facilities in China (and often elsewhere) do not enable effective monitoring and thus optimisation of temperature control or any other relevant environmental factors. Lack of optimal housing conditions is not only a bird health and welfare issue, but also a source of very significant cost to the Chinese poultry in terms of lost productivity.
Rearing chickens in sub-optimal conditions means that weight for weight, such birds require more feed than if provided with a more optimal environment. Even a very small reduction in feed can have a significant positive effect on productivity. Better conditions will also reduce the risk of disease and mortality to the birds. A significant opportunity therefore exists for UK based organisations, with expertise in this field, to collaborate with Chinese partners to develop an appropriate solution, which will in particular effect an increase in productivity within the poultry sector in China. Furthermore, it will also be a step forward in addressing the currently poor reputation of the poultry industry in China. The solution developed will have applicability in other countries around the world and may also have a use in other livestock industries. Therefore, it is an example of an excellent export opportunity for both the UK and China.
Currently poultry facilities worldwide are constructed with a few wired sensors giving a basic temperature profile in the poultry house. What is proposed is for an autonomous robot to move around the facility capturing data continuously as it moves around and sampling the environment at various heights within the facility, from the lowest cage to the highest cage. Due to the navigation knowledge of the robot it will be able to plot the data giving a 3 dimensional view of all the parameters that the birds are exposed to, to include temperature, humidity, air velocity, CO2 and will also have an on-board camera to allow visual assessment of the conditions and potentially the birds.
In addition the development of a sensor system which is based on a current prototype "e-nose" will allow data regarding the presence of disease to be captured for the first time.
Combining all the data gathered will allow assessments to be carried out that not only detects diseases before symptoms are seen in the birds, but by cross referencing all the environmental data will allow for predictive modelling to occur and potentially for the first time allow the Chinese poultry farmer to prevent sickness in their birds, improving bird welfare, bird performance and go some way to re-building consumer confidence in the product.
One key partner in the business case for this project is Applied Poultry in the UK, which already analyse data from global poultry businesses and turning it into better knowledge and information.
Once the project is completed the aim is for collaboration to continue with some data analysis being conducted by Applied Poultry. Robotics support will come from Ross Robotics. The Chinese farmers will have a dashboard for live real time analysis of their facilities and some knowledge sharing will allow the academic partners (UK and/or China) to remain involved and gaining future information from such a project.
As this proposal is for monitoring birds in cages we can see no reason why the developed robots/sensors/dashboard/data analysis cannot be considered appropriate for caged laying hens globally, although a different disease sensor arrays may be required

Planned Impact

The project addresses the challenge 1 : precision agriculture agriculture digitisation and decision management tools and falls within BBSRC strategic research priority areas: (1) Animal health (developing strategies to combat disease) (2) Sustainably enhancing agricultural production (improving survival/longevity) and (3) Welfare of managed animals (early detection of disease). Outputs include an autonomous roving platform for environmental, health and welfare monitoring and new sensor arrays to detect the digital finger print of Volatile Organic components associated with outbreaks of coccidiosis in broiler flocks. Further more a "Dashbord" collating all information from sensors on the robot and the building management system will greatly improve the farm managers ability to maximise the performance of his flock. Implementation will have an immediate relevance to poultry production and welfare. Outcomes will assist in increasing UK and Chinese competitiveness in the global livestock production market, improving animal welfare and helping to guarantee a secure supply of safe, healthy food. The following stakeholders will benefit from impact arising from this work.

1. The poultry production industry (China & UK)
Chicken production and welfare benefit from the implementation of precision livestock farming systems, but more cost effective applications of existing and novel systems are needed to increase the uptake within the tight economic margins inherent to the poultry industry. Implementation of disease monitoring using automated environmental and disease monitoring will utilise existing technology to improve animal health and welfare and economic performance. Outcomes will also be relevant to other diseases and poultry. Close relationship of the partners with the broiler industry and precision livestock technology providers will ensure the route into commercial poultry production and effective on farm use.

2. Animal welfare
The effective reduction of disease as a result of improved early detection of poultry diseases, supports the Five Freedoms implicit to animal welfare as set out by the Farm Animal Welfare Council. Further more the optimisation of the internal climate will directly benefit poultry produced.

3. General public and the environment
Increased efficiency in poultry production will raise poultry product availability at a lower cost for the consumer, contributing to improved food security. Early detection of for instance coccidiosis and targeted medication and remediation of the litter condition is likely to reduce drug consumption, the risk of contamination entering the food chain and the environment, and selection for drug resistance.
All investigators are actively engaged in public dissemination of UK research.

4. Skills, knowledge and training
The multidisciplinary nature of this project, spanning ethology, robot technology, bio-engineering and environmental science will provide opportunities for broad training to all staff, in addition to other members and students of each host institution, strengthening the research community in the areas of precision livestock farming and disease control.

Publications

10 25 50
publication icon
Fang Peng (2021) Behavior Recognition Model of Stacked-cage Layers Based on Knowledge Distillation in Transactions of the Chinese Society for Agricultural Machinery

publication icon
FANG Peng (2020) Instance Segmentation of Broiler Image Based on Attention Mechanism and Deformable Convolution in Instance Segmentation of Broiler Image Based on Attention Mechanism and Deformable Convolution

 
Description Using an specialist electronic sensor array (electronic nose) for the detection of volatile organic compounds (VOCs) we are able to detect the presence of "Coccidiosis" a disease afflicting poultry and broilers in particular. Coccidiosis has severe negative broiler welfare implications (ultimately death) as well as significant negative impact on broiler production parameters. To date the results indicate that coccidiosis is identified at least 48hrs earlier than the first clinical signs are visible to the stockman/veterinarian giving increase scope to treat the disease early and prevent to a large extend its negative impact on bird welfare and production.
A prototype autonomous roving platform and sensor array was designed, build and tested in clean conditions in the UK for this project. However, due to funding cuts and covid19 related delays the robotics partner in the project was forced to withdraw. An alternative simplified version was build in China and is now used on farm trials in China. This allows for monitoring of the internal climate of the full building as opposed to the current practise of monitoring and controlling the internal climate with a few censors near the centre of the building. The farm managers are aware there are large localised differences in indoor climate quality over time. Using the information collected using the robot our data analysts are able to provide real time advice to make improvements to the internal climate.
Exploitation Route This award is still active.
The early disease detection with the VOC sensor is likely to lead to improved disease management on farms where the system is implemented, preferably on a autonomous moving platform as currently under development in this project, but also as a standalone health monitoring system. This will have a significant impact on bird health and welfare and potentially reduce the use of antibiotics in the industry providing the consumer with an affordable healthy source of protein.
• The VOC sensor will help the poultry industry to provide a healthy, safe and affordable source of protein (world wide, not just ODA coutries)
• Reduction in antibiotic use, important especially in ODA countries in Asia, where antibiotic use is relatively high and drives up the cost of the end product.

Due to delays caused by covid 19 and the cuts in the funding to the project, the completed autonomous roving platform development prototype is no longer available to the project team due to the withdrawal of Ross Robotics Ltd. However, large elements of the development are being taken forward by Ross Robotics in other commercially sensitive projects (Charging module, main navigation module, avoiding "ghost" objects detected by the Lidar system, ....).
• Two autonomous roving platforms have been developed (UK & China) providing opportunities for Chinese partners to develop the system to a commercially viable end product

Due to the delays the developed end user Dashboard has not yet been used in this project. However, a version is already used commercially by OptiFarm for their end of crop reporting. This includes several farms in China and Asia.

As the project is finally entered a the on farm testing phase of the robot and associated technology (reduded to just two months), we can not yet attribute the further outcomes to the associated outputs of the project.
Sectors Agriculture, Food and Drink

 
Description Visual Monitoring of Broiler Behaviour, Health and Welfare using Artificial Intelligent Image Machine Learning
Amount £188,986 (GBP)
Funding ID 10006622 
Organisation Innovate UK 
Sector Public
Country United Kingdom
Start 11/2021 
End 04/2023
 
Description Collaboration with project partners in China 
Organisation China Agricultural University (CAU)
Country China 
Sector Academic/University 
PI Contribution Intensive collaboration between TOBOR, Chinese Agricultural University, Royal Veterinary College and Ross Robotics in the development of the datalogging module hardware, software and software integration within the project. the RVC lead specification and selection of the sensors used and contributed to the software development and testing. Acquisition and testing of sensors. RoboScientific and RVC have trialled and validated the Volatile Organic Component sensor array for detection of Coccidiosis spp in broilers in ongoing vaccine development trials at the RVC.
Collaborator Contribution all collaborators are partners in the project. TOBOR build the environmental and camera data logging modules and developed the software for the modules. Ross Robotics designed and build the UK prototype robot, set the parameters for integration of the software with the Robot control and navigation systems ; developed the application programming Interface (AP)I for the integration and provided assistance in the software development. CAU contributed to the selection and acquisition of sensors. CAU build the software for data transfer of the modules and the database. CAU has also build the alternative robot to be used on the Chinese farms. Applied Poultry developed the farmer facing "Dashboard". RVC set the specifications for the data transfer, data base, coordinated the development and involved in the testing of the hardware/software. RoboScientific have significantly improved the sensor array (selection and production) used in the VOC sensor. Secondly, a new sampling, absorption and desorption system was designed, tested and after several iterations proven to be suitable.
Impact RoboScientific has patented elements of the VOC sensor based on the work done within this project. Still ongoing
Start Year 2019
 
Description Collaboration with project partners in China 
Organisation Hudson & Saunders Limited
Country United Kingdom 
Sector Private 
PI Contribution Intensive collaboration between TOBOR, Chinese Agricultural University, Royal Veterinary College and Ross Robotics in the development of the datalogging module hardware, software and software integration within the project. the RVC lead specification and selection of the sensors used and contributed to the software development and testing. Acquisition and testing of sensors. RoboScientific and RVC have trialled and validated the Volatile Organic Component sensor array for detection of Coccidiosis spp in broilers in ongoing vaccine development trials at the RVC.
Collaborator Contribution all collaborators are partners in the project. TOBOR build the environmental and camera data logging modules and developed the software for the modules. Ross Robotics designed and build the UK prototype robot, set the parameters for integration of the software with the Robot control and navigation systems ; developed the application programming Interface (AP)I for the integration and provided assistance in the software development. CAU contributed to the selection and acquisition of sensors. CAU build the software for data transfer of the modules and the database. CAU has also build the alternative robot to be used on the Chinese farms. Applied Poultry developed the farmer facing "Dashboard". RVC set the specifications for the data transfer, data base, coordinated the development and involved in the testing of the hardware/software. RoboScientific have significantly improved the sensor array (selection and production) used in the VOC sensor. Secondly, a new sampling, absorption and desorption system was designed, tested and after several iterations proven to be suitable.
Impact RoboScientific has patented elements of the VOC sensor based on the work done within this project. Still ongoing
Start Year 2019
 
Description Collaboration with project partners in China 
Organisation RoboScientific
Country United Kingdom 
Sector Private 
PI Contribution Intensive collaboration between TOBOR, Chinese Agricultural University, Royal Veterinary College and Ross Robotics in the development of the datalogging module hardware, software and software integration within the project. the RVC lead specification and selection of the sensors used and contributed to the software development and testing. Acquisition and testing of sensors. RoboScientific and RVC have trialled and validated the Volatile Organic Component sensor array for detection of Coccidiosis spp in broilers in ongoing vaccine development trials at the RVC.
Collaborator Contribution all collaborators are partners in the project. TOBOR build the environmental and camera data logging modules and developed the software for the modules. Ross Robotics designed and build the UK prototype robot, set the parameters for integration of the software with the Robot control and navigation systems ; developed the application programming Interface (AP)I for the integration and provided assistance in the software development. CAU contributed to the selection and acquisition of sensors. CAU build the software for data transfer of the modules and the database. CAU has also build the alternative robot to be used on the Chinese farms. Applied Poultry developed the farmer facing "Dashboard". RVC set the specifications for the data transfer, data base, coordinated the development and involved in the testing of the hardware/software. RoboScientific have significantly improved the sensor array (selection and production) used in the VOC sensor. Secondly, a new sampling, absorption and desorption system was designed, tested and after several iterations proven to be suitable.
Impact RoboScientific has patented elements of the VOC sensor based on the work done within this project. Still ongoing
Start Year 2019
 
Description Collaboration with project partners in China 
Organisation Ross Robotics
Country United Kingdom 
Sector Private 
PI Contribution Intensive collaboration between TOBOR, Chinese Agricultural University, Royal Veterinary College and Ross Robotics in the development of the datalogging module hardware, software and software integration within the project. the RVC lead specification and selection of the sensors used and contributed to the software development and testing. Acquisition and testing of sensors. RoboScientific and RVC have trialled and validated the Volatile Organic Component sensor array for detection of Coccidiosis spp in broilers in ongoing vaccine development trials at the RVC.
Collaborator Contribution all collaborators are partners in the project. TOBOR build the environmental and camera data logging modules and developed the software for the modules. Ross Robotics designed and build the UK prototype robot, set the parameters for integration of the software with the Robot control and navigation systems ; developed the application programming Interface (AP)I for the integration and provided assistance in the software development. CAU contributed to the selection and acquisition of sensors. CAU build the software for data transfer of the modules and the database. CAU has also build the alternative robot to be used on the Chinese farms. Applied Poultry developed the farmer facing "Dashboard". RVC set the specifications for the data transfer, data base, coordinated the development and involved in the testing of the hardware/software. RoboScientific have significantly improved the sensor array (selection and production) used in the VOC sensor. Secondly, a new sampling, absorption and desorption system was designed, tested and after several iterations proven to be suitable.
Impact RoboScientific has patented elements of the VOC sensor based on the work done within this project. Still ongoing
Start Year 2019
 
Description Collaboration with project partners in China 
Organisation TOBOR Technologies
Country United States 
Sector Private 
PI Contribution Intensive collaboration between TOBOR, Chinese Agricultural University, Royal Veterinary College and Ross Robotics in the development of the datalogging module hardware, software and software integration within the project. the RVC lead specification and selection of the sensors used and contributed to the software development and testing. Acquisition and testing of sensors. RoboScientific and RVC have trialled and validated the Volatile Organic Component sensor array for detection of Coccidiosis spp in broilers in ongoing vaccine development trials at the RVC.
Collaborator Contribution all collaborators are partners in the project. TOBOR build the environmental and camera data logging modules and developed the software for the modules. Ross Robotics designed and build the UK prototype robot, set the parameters for integration of the software with the Robot control and navigation systems ; developed the application programming Interface (AP)I for the integration and provided assistance in the software development. CAU contributed to the selection and acquisition of sensors. CAU build the software for data transfer of the modules and the database. CAU has also build the alternative robot to be used on the Chinese farms. Applied Poultry developed the farmer facing "Dashboard". RVC set the specifications for the data transfer, data base, coordinated the development and involved in the testing of the hardware/software. RoboScientific have significantly improved the sensor array (selection and production) used in the VOC sensor. Secondly, a new sampling, absorption and desorption system was designed, tested and after several iterations proven to be suitable.
Impact RoboScientific has patented elements of the VOC sensor based on the work done within this project. Still ongoing
Start Year 2019
 
Title Automatic Chicken house monitor 
Description The technology has the ability to inform farmers when an infectious disease occurs in a flock of chickens, enabling faster interventions, reducing wastage and increasing yields All organic bodies, animal or vegetable, emit odours or Volatile Organic Compounds (VOCs). Any change to the chemistry caused by disease or impurities will change the balance of VOCs emitted and this will be detected by our sensitive semiconducting polymer sensors. Powerful software then takes this data & analyses the odour "digital fingerprint". Comparison with stored fingerprints enables diseases and other problems to be quickly detected. In this project the automatic chicken house monitor was developed, especially the integration of the automated sampling and analysis system in one device, including the bespoke analysis and instrument control software. Furthermore the sensor array was adapted to include detection of the onset of the disease Coccidiosis by the instrument 
Type Of Technology Detection Devices 
Year Produced 2022 
Impact Product is on sale for use by farmers. 
URL http://www.roboscientific.com/products/poultry/
 
Title dashboard for farmer use/communication 
Description the "dashboard" Developed by Hudson and Saunders (Applied Poultry) is the interface between the management consultants at OptiFarm and the individual farmer/farm manager, providing the latter with the latest recommendations for optimal farm management. this new version incorporates bird health and welfare sections based on the data provided by the Envirobot program and existing produciton parameters (feed/water consumption, temperature etc) 
Type Of Technology Software 
Year Produced 2020 
Impact Better bird health and welfare are likely to reduce production costs and will lead to a widely available, healthy, low cost source of animal protein which is of essence in DAC countries (including large parts of China). 
 
Description presentatio/discussion re robotics in poultry production 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact presentation at Defra (Welfare section and infectious disease experts) by Dr Theo Demmers (RVC), Philip Norman (Ross Robotics) and David Speller (Applied Poultry) re use of autonomous robotic monitoring platforms in poultry production (Broilers and Laying Chickens, as well as breeding stock). Discussion re potential use of commercial products by stakeholders for various purposes.
Year(s) Of Engagement Activity 2020