Combining AR Technologies and Electronic Animal Identification for Digital Livestock Farming

By Samudrapom DamReviewed by Susha Cheriyedath, M.Sc.Oct 27 2023

In an article recently published in the journal Scientific Reports, researchers developed the SmartGlove framework and investigated the feasibility of using the framework to effectively connect animal electronic identification (EID) with augmented reality smart glasses (ARSGs) for the visualization of specific animal data directly in the field, providing a hands-free environment.

Background

Several technologies, such as radio frequency identification (RFID), have been introduced in recent decades in the livestock farming sector. RFID systems, primarily composed of a transponder and a transceiver/reader, are utilized for the EID of animals.

In modern farms, different types of data are obtained from various sensors due to the proliferation of precision livestock farming (PLF) technologies. PLF sensors can be wearable by the animal, such as ear tags, collars, and bolus, or fixed, such as weather stations and cameras.

Although PLF technologies can provide a substantial amount of data, the consultation and interpretation of these data by farmers is a time-consuming and difficult task. Visualizing the summarized or raw data directly on-field is crucial for effectively utilizing sensor outcomes. However, the data consultation process can halt normal farm management activities even when the process is performed through mobile devices, such as a tablet or smartphone, as these devices must be handheld by the operators.

In such scenarios, ARSGs connected to EID can allow data visualization/consulting specific animal databases directly on-field to offer a hands-free environment for the operator. ARSGs are primarily wearable head-up displays that are connected to/integrated with a miniaturized computer, which adds virtual information to the reality of the user.

The SmartGlove framework

In this study, researchers developed and evaluated a wearable framework, designated as SmartGlove, which can link RFID animal tags and ARSGs through a Bluetooth connection to allow the visualization of specific animal data directly in the field. The objective of the study was to develop a framework that allows a connection between animal electronic identification and various ARSG types and to evaluate the framework performance on-field and in the laboratory.

Researchers also performed a comparative analysis of different AR technologies, including mixed and assisted reality systems, to determine the most effective solution for livestock farm environments. SmartGlove is a technology readiness level-3 (TRL-3) prototype that allows reading unique animal code from RFID tags and sending the code to ARSG to display all relevant information related to the specific animal.

The prototype hardware consists of an RFID reader board connected to a 125 kHz antenna and an Arduino control unit with integrated Bluetooth. All components are enclosed in a three-dimensional (3D) printed plastic case, which can be utilized as a bracelet, with the 125 kHz copper antenna extending to the back of the hand and attaching to a glove.

The SmartGlove was connected to two ARSG types, including the Epson Moverio BT–300 (BT300), which is an Android-based assisted reality device with a binocular, optical, see-through display, and the Microsoft HoloLens 2 (HL), which is a Microsoft Windows-based mixed reality device with a holographic display.

Android software development kit (SDK) and Universal Windows Platform (UWP) were used for BT-300 and HL devices, respectively, to develop and design supporting applications that run on these devices. The software solution comprised three modules, including the headset application, the database, and the glove hardware manager.

Experimental evaluation and findings

Researchers designed two sets of trials to assess the operational performance of different systems. The first set was conducted in the laboratory to evaluate the complete operativity of SmartGlove and assess its RFID tag reading and ARSG connection performance in a controlled environment. All tests were performed using two types of tags, including the full-duplex (FDX) ear tag (ET) and FDX rumen bolus (RB), and with both BT300 and HL ARSG.

The second set was performed at the experimental sheep farm of the University of Sassari to investigate the developed framework’s operating capabilities in a real farm scenario. A commercial RFID reader/F1 reader was employed to compare the developed systems during all tests. Additionally, the point of view of stakeholders/feedback of participants on the developed system was analyzed using two standard questionnaires, including the IBM-Post Study System Usability Questionnaire and the NASA-Task Load Index (TLX) questionnaire.

In laboratory tests, the success rate of the commercial RFID reader and the SmartGlove with both ET and RB transponder types varied based on the activation distance. The SmartGlove framework displayed an acceptable success rate of 70% for ET and 60% for RB transponders at an activation distance of three cm, and a 100% success rate for both transponder types at one cm activation distance.

The F1 reader demonstrated an acceptable success rate of 65% for RB and 95% for ET at a five cm activation distance and a 100% success rate for both transponders at two cm. HL demonstrated a higher variability in the reading process time compared to BT300. The difference between the minimum and maximum time was 7.90 s and 2.87 s for HL and BT300, respectively.

In the on-field tests, participants performed the animal identification in a shorter time using the conventional system/F1 reader+paper list compared to the assisted reality smart glass (AaRSG) system/BT300+SmartGlove or mixed reality smart glass (MRSG) system/HL+SmartGlove.

Similar results were obtained for average time per tag reading, with participants completing their task much faster using the conventional method compared to AaRSG and MRSG systems. Among the augmented reality systems, AaRSG displayed a better performance compared to MRSG in on-field tests. No error was observed in all tested systems.

However, the participants’ feedback confirmed a high usability level and a low cognitive impact for the use of the SmartGlove connected to ARSG, with 11 out of 18 participants selecting HL+SmartGlove and seven selecting BT300+SmartGlove as the most preferred system. No participant selected the conventional method, with 12 out of 18 participants indicating it as the least preferred system.