Dr. Patrick Jackman patrick.jackman1@ucd.ie
We are a research group based in University College Dublin Biosystems Engineering We specialise in Big Data solutions to real world Agricultural and Food Engineering problems www.ssu.ie We currently work for multiple SME s, Industrial partners and Government Agencies Prof. Shane Ward: Principal Investigator Dr. Liam Brennan: Lecturer Dr. Gerard Corkery: Lecturer Dr. Ultan McCarthy: Technology Gateway Manager: IMaR Dr. Patrick Jackman: Research Scientist Dr. Maeve Cushen: Research Associate Prof. Colm O Donnell: Senior Lecturer Dr. Kevin McDonnell: Senior Lecturer Dr. Tom Curran: Senior Lecturer
Global broiler production market stands at approximately 100million tonnes of meat This is expected to grow significantly into the next decade Market growth is expected to be particular strong in the emerging BRIC markets
European market is massive at over 10 million tonnes per annum While production is massive profit margins are tiny Thus a small gain in productivity can lead to a huge gain in profitability Small gains could also turn some failing businesses into viable ones Poultry share of the total meat market has grown from 22% to 28% and is expected to grow to 31% EBITDA MARGIN TREND 2002-2012
Strategy for making marginal gains comes from the concept of Precision Livestock Farming This involves tight control of every variable in the production and process chain By eliminating the small losses a noticeable overall gain will be achieved Essential features of precision livestock farming are: Automation of measurement Automation of measurement interpretation Diagnosis of control breaches Local automatic control systems Increasing network of regulatory requirements adds further need for Precision Livestock Farming
Precision Livestock Farming will create enormous datasets that must be interpreted Data interpretation leads to a diagnosis of any problem that might exist Predictive model decides the appropriate remedial and mitigating action Local automatic control implements the prediction to quickly eliminate the problem Thus production and process efficiency is maximised leading to yield maximisation
Big data also offers opportunity for forensic analysis of production and process chain Any interested party can probe the big data for their information of interest Wholesalers can use the big data to monitor process production Retailers can use the big data to monitor their wholesale supply Consumers can use the big data to satisfy themselves of the traceability of the products they buy off retailers
The growing environment is essential for poultry yield as comfortable chickens put on a lot of weight High levels of gases such as CO 2 and NH 3 can lead to poor quality yields e.g. disease and foot ailments Impact: Early identification of loss of process control
The Bosca project places wireless sensors in the chickens living space to record temperature, humidity, air speed and gas concentrations Real time spatial and temporal data is reported to a cloud repository via 3G technology Impact: Real Time consolidation of observed data
The Bosca data allows deviations from acceptable environmental limits to be quickly detected and quickly mitigated by the producer A spatial and temporal map of poultry houses can be generated from the Bosca data as each Bosca has a GPS location and is aware of Unix Time The spatial and temporal map can be compared to productivity data so wholesalers can identify their best and worst poultry producers Impact: Forensic identification of chronic process problems
The Chirpmetrics project will add another independent source of chicken contentment data If chickens are content they will coo in particular ways If chickens are discontented they will coo in other ways Features extracted from acoustic time series are added to the data cloud Impact: Additional supporting Real Time data recorded
The Talking Truck places a network of environmental sensors into a poultry delivery truck Uploads data to cloud repository in real time and produces a spatial and temporal map as sensors have GPS location and are aware of Unix Time Ensures that poultry are correctly handled during transit and that yield and welfare are protected ENVIRONMENT POWER SOURCE PRODUCT DATA TRANSFER TAG TYPE POWER PALLET ORIENTATION Impact: Identification of potentially damaging product mishandling
The CyberBar project provides a tracking mechanism from when the chicken is slaughtered in the processing hall to the Supermarket shelf The history of the chickens that are loaded into the slaughter house process queue is known from Bosca, Chirpmetrics and Talking Truck Immediately post slaughter they are branded with a unique Quick Read Code Impact: Instant recognition of product history providing full traceability
The unique Quick Read code can be read by any device with a camera and a Quick Read app such as a smart phone The full history of chicken fillets can then be immediately retrieved from the cloud data repository Thus the retailer can gain quality and animal welfare assurances from their wholesaler Similarly a customer can hold their retailer to account as complete information on chicken farming, handling, transport and slaughter is associated with the Quick Read code Impact: Instant recognition of product history providing full traceability
The entire Big Data repository can be queried graphically via a web portal or in tabular form by SQL or SPARQL endpoints by any interested party As all data entries have a GPS and Unix Time almost every conceivable query is possible and visually appealing Google Maps can be generated Impact: Instant transmission of important traceability parameters
There is no reason why the Big Data concept cannot be re-applied to any production and processing cycle Pig and Beef production are obvious next targets for the Big Data solution Data recording sensors would need some modifications and improvements but nothing radical Cloud repository would similarly need some tweaking and refinement As long as the data consumption requirements do not greatly change the hardware and software can be adapted to almost any challenge where GPS and Unix Time are important
The Smart Sensors Unit will be leading a Big Data in Agriculture proposal for 2015 call of EU Horizon 2020 Proposal will design a Smart Crop system that will provide comprehensive information on every aspect of monitoring mycotoxins in crops pre and post harvest By providing Real Time Big Data and a Cloud based solution, efficient and effective infield management and storage management of mycotoxins becomes a reality The Smart Sensors Unit will be seeking partners and collaborators for discussions on forming a consortium to form the project bid Contact: patrick.jackman1@ucd.ie
Big Data from sensors aware of GPS and Unix Time provides a comprehensive spatial and temporal map Cloud computing offers the framework for housing the enormous datasets A well designed data cloud data can be easily queried for the data of interest both graphically and in tabular form Solves key accountability and traceability questions for wholesalers, retailers and consumers who wish to keep control on their suppliers Increased productivity, profitability and compliance will result from the application of Big Data solutions