ABSTRACTA lot of research has been done in the field of wearables for animals in the last time. Wearables can not only be used as an entertainment or communication tool between humans and pets, but they also support health management of different animals, preventing in this way the progression of different health issues into worse stages. This paper focuses on the sensor systems that support husbandry in managing the health of dairy animals. It provides an overview of technologies and applications developed for cattle, referring to industry and research work. The sensors are firstly categorized into groups, based on the position where they can be placed. Afterwards, the paper describes the most used sensors for mastitis, fertility, locomotion and metabolism problems, considered in past research as the most costly health issues of dairy animals.Author Keywordssensor systems, wearables, dairy, health management, cattle, mastitis, locomotion, metabolism, fertilityACM Classification Keywordssensor devices, health informaticsINTRODUCTION We are living in a digital world, where digital devices are increasingly affecting our lives, by handling different tasks and making our lives easier. Wearables have also been a part of the digitalization trend and process in the last years, not only for humans but for animals as well. The automation of processes can especially help dairy farmers to reduce physical labour and labour costs, considering the increase of farm sizes and the need of higher-yielding cattle. 3 This review delivers an organised overview of the wearable sensors developed to assist with four different health issues of dairy animals like mastitis, fertility, locomotion and metabolism. This review is based only on the sensors developed for these issues because, in this order, these are the four most costly health problems of dairy animals. 5DEFINITION AND CLASSIFICATIONA sensor can be defined as a device which measures a physiological or behavioural parameter of an animal, in this case of a dairy animal. It enables automated detection of changes in a condition that require action from the farmer or the vet. A sensor system consists of the device itself and the software processing the gathered data. 15Sensors can be divided into two categories: wearables/attached and non-wearables/ nonattached. Wearables can be divided into two subcategories which are on-cow and in-cow sensors. The on-cow sensors can be put outside the body of the cows like pedometers and accelerometers whereas the in-cow sensors are devices which can be inserted inside the body of the animal, for example, rumen bolus or implants. 15On the other side, non-wearables are off-cow devices where the cow can pass by, over or through for different measurements. The two subcategories of non-wearables are in-line and on-line sensors. There is a remarkable difference between these two types of sensors. In-line sensors take measurements in a continuous flow of a product. Hence, the only product of dairy animals that can be measured by an in-line sensor is the milk flow. These sensors are mostly attached to Automated Milking Systems (AMS) where the continuous flow of the milk is possible. A good example is the Afimilk milk analysis technology, which will be discussed in the next paragraphs of this overview. On the contrary, on-line sensors are based on sampling. Thus, the product samples can be taken to a laboratory where different features can be measured but not in real time. 15SELECTED SENSORSThis overview discusses the different sensors developed for the four most costly health problems of dairy animals: mastitis, fertility, locomotion and metabolism. 5 Each of the following paragraphs provides a short overview of the statistics for each category and then describes the most studied and used sensor for the specific disease.MastitisStarting with the first disease, mastitis is defined as an inflammatory reaction in the mammary gland. The inflammation in the udder tissue may be the result of bacterial, chemical, thermal or mechanical injury. It is the most common and most costly disease of dairy cattle. 10 There are found 37 sensor systems which can detect mastitis. Half of them use electrical conductivity and the second most used ones are biosensors which detect different enzymes. 15Mammary infections increase the somatic cell count in ruminants. Somatic cell count is a common method for the evaluation of milk quality and of mammary gland health status in dairy animals. 11 Mastitis modifies the concentration of anions and cations and can affect milkelectrical conductivity, so the evaluation of electricalconductivity represents an interesting approach. 21Another widely used technology is the milk analyzer. Milkcomposition can provide information about nutrition,metabolic state and health through changes in the milkcomponents. 8 A good example is the ‘AfiLlab’ systemdeveloped from Afimilk’s which combines the somatic cellcounter with measurements of milk components (i.e. fat,protein and lactose) and milk coagulation. 9FertilityThe second health issue that is being contemplated isfertility. 61% of the sensors study the activity of the cow,60 % of them give an estrus alert, but only 15% of themshow the probability of the estrus alert as well. 15 Theobjective of sensors developed for fertility issues is toprevent missing cases of true estrus caused by using visualobservation. 6 The technologies studying the activity ofthe cow during heat periods are mostly pedometers andaccelerometers.The oldest and most studied sensor is the pedometer. 19Pedometers count each step the animal takes by detectingthe motion of its hips and legs. Dairy animals take a highernumber of steps during the estrous period, especially whenmore animals are in behavioural estrous at the same time.Furthermore, different researches show that ovulationoccurs 19.4 ± 4.4 hours after the end of an increasednumber of steps. Considering this behaviour feature ofdairy animals, pedometers detect estrous accurately and canimprove fertilization rates. 14Accelerometers are used in a similar way but they recordinstead three-dimensional movement and can be used tosubstitute camera surveillance. They gather data about thebehaviour of the cattle, if it is standing, sitting, lying ormounting. 13 There are already different developments inthe industry, for example, ‘SensOor’,’Cow Alert’,’CowScout’, ‘Qwes’, ‘Activity meter’ etc. 1LocomotionLameness, the locomotion problem of cattle, is consideredto be the third most influential health issue, after mastitisand fertility issues. Researchers have observed thefrequency of this disease in 21 farms and the physicaleffects caused by it on 2183 cows. The final result showedthat a lame cow causes a loss of 110 € per foot per year. 545 % of the sensors developed to detect lameness study theweight distribution between the legs, whereas 55 % of themstudy their walking behaviour and activity. 15 Pedometersand 3D accelerometers are the most studied sensors. 4 Asthe way these sensors work is described in the paragraphabove, it is interesting to see the sensors that measure theweight distribution between the legs as well. According to astudy and experiment conducted by Pastell M. et al.,measurement of how cows distribute their weight whenstanding holds promise as a method of automated detectionof lameness. 12 There are already two products in theindustry that are using gait consistency analysis and forcesensors for lameness detection: ‘StepMetrix’ and’Gaitwise’. 1 These sensors evaluate and analyze theforce, pressure and duration of each step when the cows arewalking across the Step Sensor platform after each milking.Comparing them with the gold standards of the system, theycan easily track lameness even on the whole herd. 17MetabolismKetosis is the most common metabolism disease ofruminants. Ketosis is defined as an increase of ketonebodies (chemicals that break down fat instead of the sugarglucose for energy) found in blood, urine, or milk. 7 38%of the sensors developed are radiotelemetric boluses thatmeasure the pH of rumen fluid and temperature. 15 Someof the reticulo-rumen boluses developed that are already inthe animals wearables market are ‘eCow’, ‘WellCow’,’smaXtec pH’, which are inserted orally and can assessrumen function by temperature and acidity by pH, and theirinformation helps the farmers detect digestive problems atan early stage and prevent further costs andcomplications.1Milk analysis technologies and metabolic biomarkers,which are used for mastitis, can also be used for thedetection of ketosis. A good example is the Herd Navigatorfrom Lattec. The Herd Navigator system not onlyautomatically selects the cow to sample and the milkingsession but it also selects the variable of the milk thatshould be measured. Ketosis can be detected through theanalysis of a chemical substance named ?-hydroxybutyratefound in the milk.CONCLUSIONIn conclusion, this work describes the most used andimportant wearables used for cattle health management.However, the development of wearables for dairy animalsis still in its early stages. There has been more research anddevelopment done on sensors for mastitis and fertilityissues, whereas sensors helping with locomotion andketosis detection are still in the technique developmentphase. Therefore, the alerts coming from the wearables formastitis and estrus are more informative than thelocomotion alerts. 15 Based on the findings of thisliterature overview, we can observe that this topic is stillquite innovative and there is still a lot of room forimprovements, especially for detecting locomotion andmetabolism problems; new developments which canintegrate information from different systems and forresearch, especially in the field of economic profitability ofdetecting and treating the diseases in time.