Design of an intra-ruminal wireless communication sensing system for measuring temperature and pH of cattle

The use of mechanical and electronic devices in different livestock production processes has led to automation of a set of practices. Although these practices have initially reduced the need for labor, their main contribution is the possibility of adapting operations to individual needs of the animals, for which there must be subsystems capable of recognizing animals when they interact with automated systems (Bijl et al., 2007; Cornou, 2009; Betteridge et al., 2010; Hamadani et al., 2015). Biosensors have the ability to measure physiological and immunological responses in various animal species based on the generation of new biosensing methodologies. To achieve this, specific monitoring equipment is available, with a high degree of specialization that covers several aspects of animal physiology and environment. In addition, these systems meet desirable conditions such as reliability, ease of use and the opportunity to improve the monitoring of desired variables (Bijl et al., 2007; Betteridge et al., 2010; Sinha et al., 2018).

In cattle, these biosensors have allowed evaluation of numerous physiological and metabolic variables that have been commonly associated with productive processes (Sinha et al., 2018) and ruminant health, such as ruminal stability. Measurement of pH, whose variations determine rumen functionality, has allowed establishing a range of pH, variations outside of which animal health can be affected (Duffield et al., 2004; Castro et al., 2015; Neethirajan et al., 2017). A common problem in dairy cattle is ruminal acidosis, characterized by an alteration of pH when the amount of saliva produced to buffer the pH of the rumen decreases and the ingestion of large quantity of carbohydrate rich diet, causes high mortality due to proliferation of lactate producing rumen bacteria (Kumar et al., 2007), pH can also be altered by consumption of feed concentrate in larger quantities or sudden changes in the diet.  Besides voluntary intake (Duffield et al., 2004; Castro et al., 2015), thermal environment is a major factor that can negatively influence animal physiology especially of high genetic merit animal (Yazgan, 2017), comfort depends on environment temperature. In the case of rumen temperature, detection can be adjusted to between 37°C and 40°C. A biosensor that senses this range covers all relevant physiological temperatures and thus, it is useful for determination of the main ruminal functions. It would be of use, for example, to physiologists interested in the complex ruminal functioning and to veterinarians responsible for animal welfare (Duffield et al., 2004; Castro et al., 2015; Tousova et al., 2017).

Animal welfare means how an animal is facing the conditions in which it lives; it is based on ‘five freedoms’ and ‘four principles’ of animal welfare. Dairy cattle are considered as sentient beings due to which husbandry should be provided as per their needs (Kumar et al., 2017) which are influenced by factors such as temperature and relative humidity, management, nutrition and preventive medicine (Moberg, 1987; href="#belasco_2015">Belasco et al., 2015; Tousova et al., 2017). The productive efficiency of the animal can indicate its relationship with the production system in which it is managed; therefore, its analysis can provide a starting point for assessing welfare (Duncan, 1990). Bovines maintained and managed under optimal conditions can better express their breeding characteristics, which economically favor production. The duration and frequency of the periods during which the cows lie down are indicators of the degree of comfort they are experiencing (Duncan, 1990; Haley et al., 2000). This has direct relevance for clinical health, particularly, in the incidence of lameness. In addition, plasma concentrations of growth hormones are reduced in cows that are deprived of rest periods and can affect milk production (Munksgaard and LØvendahl, 1993). The rest periods of bovines are observed directly by personnel in charge or recorded in video for a later analysis. These methods, however, are time-consuming, requiring intense work and may influence the behavior of the animals under study.  In addition, when more than one observer is used, inter-observer differences arise (Muller and Schrader, 2003; Neethirajan, 2017). Moreover, none of the methods mentioned for data collection is suitable for use in large-scale production systems. The development of autonomous probes for monitoring the biological variables ruminal pH and temperature with electronic elements that carry out conversion and transmission of in vivo data with high temporal resolution is the main focus of this study. This paper presents the results of the development of a multifunctional prototype capable of simultaneously measuring the two variables and sending these signals to a receiver for real-time viewing.

Experiment design
 
During the first stage, the sensing system was immersed in a microenvironment that emulated the characteristics of the cow rumen. The second stage was introduction of the system into the rumen of a cannulated cow. Care and handling of the animal complied with the official Mexican standard NOM-062-Z00-1999, according to the Bioethics Regulations of the School of Veterinary Medicine and Animal Sciences of Veracruz University.
 
Device design
 
The embedded system through which collecting, conditioning, transmitting and receiving data is controlled by an Arduino® System (Fig 1A). This system is an open-source electronic physical computing development platform, based on a board that incorporates an ATmega328 microcontroller® and an ad hoc environment for software development. The Arduino® system has some advantages over other boards. For example, its programming language is easier than C, C++, etc. and no extra hardware is needed; it consists of RAM, ROM, supply, analogue, and digital pins that are required to run an electronic system. The computer system used was a laptop with 600 GB DD, 4 GB of RAM and 64-bit Windows 7® operating system. For the storage and visualization of the data, Matlab 2013-B® was used. This microcontroller sends and receives data through a serial connection to an Xbee® module (Fig 1B), 50 mW and an UHF band at 2.4 GHz. Using these modules, a wireless network with more than 10 serial connections was developed capable of transmitting up to a distance of 1 km.

Other elements that support this design are the LM35 series and precision integrated-circuit temperature devices with an output voltage linearly-proportional to temperature (°C). The LM35 device does not require any external calibration or trimming to provide typical accuracies of ± 0.25°C at room temperature and ± 0.75°C over a full -55°C to 150°C temperature range. (Fig 1C). For pH sensing inside the rumen, a Sensorex® SG201C Glass Body Bulb-Style Combination pH electrode was used. This electrode has a 0-14 pH measurement range, operating temperature range 0-100°C, glass body, diameter 12 mm, length 150 mm, and reference Ag/AgCl (Fig 1D).



The impedance of the pH electrode is in the range of 10 MW to 100 MW, and then a stage of signal conditioning supported by a voltage follower circuit based on a TL084 Texas Instruments JFET-input operational amplifier family was designed for impedance matching. As a second stage, a first order low pass active filter (fc=0.0328 Hz), was incorporated with the purpose of reducing the interference of spurious signals. From the technical specifications of the glass electrode, the equipotential point is at pH 7.00 (0 mV). Next, a calibration circuit based on a subtractor/difference operational amplifier stage was connected to the previous circuit from which a reference signal goes to an input terminal and the electrode’s signal enters the other input terminal. This configuration amplifies the difference between the two signals. At the output node of the differential stage, the signal range is between±150 mV, which is not suitable for the next digital conversion process, and thus we sought to work in a voltage range of 0-4V. For this purpose, a stage integrated by an op-amp inverter (gain=13.3) circuit and an op-amp inverting adder unit gain circuit is built (Fig 2). Once the signal coming from the pH sensor is processed in the conditioning stage, it is sent to the Arduino®  A/D converter block.

​

Since the integrated circuit TL084 is an element used in each module of the analog signal conditioning stage of the probe signal, in the laboratory a multiple output DC power supply was used to apply±5V. However, when the prototype is inserted into the rumen of a cow, it is necessary to have a source of energy that allows the system to operate autonomously. Considering the characteristics of the process, the system employs a 3-cell lithium polymer battery (LIPO) with a capacity of 2200mAh and a voltage of 11.1V (Fig 3). This battery will supply power to the entire circuit. Once circuit design specifications were considered, a 1.5V wide-input adjustable buck-boost switching regulator, PTN78000A Texas Instrument®, was used to generate a regulated -5V power supply and the simple switcher power converter 150KHz 3A step-down voltage regulator LM259SX-5.0 GTM® to generate a regulated +5V power supply (Fig 3).

​
 
Clinical trial (in vivo)
 
The ruminal bolus is 25 cm long and 10 cm in diameter, with a net weight of 491 g (Fig 4A, B, C, D). Thermoplastic was used for fabrication of the prototype with a Tronxy X5S High-precision Metal Frame 3D Printer Kit. The ruminal bolus was designed using SolidWorks 2017™ computer-aided design. The top of the bolus (lid) was threaded to assist extraction of the prototype more easily from the bottom of the rumen. The middle part of the bolus is a compartment designed to house the electronics together with the sensors and battery. The sensors make contact with the ruminal environment through a series of holes at the bottom of the bolus (Fig 4E).

​
 
Statistical analysis
 
Data were analyzed with simple linear correlation in the GLM module of the STATISTICA v10 software, and Sigma Plot v11 was used to construct the figures.

The electronic physiological monitoring systems were validated with an adjustable temperature incubator (Isotemp 228, Fisher Scientific®) (25-100°C). The systems were placed in a 12-liter-capacity tray containing 8 liters of water and fodder (tropical grass harvested on the day of the experiment). The temperature in the incubator was obtained by means of a glass liquid thermometer (1/300°C). The measurement system was placed inside the incubator for an interval of 30 min after mixing the simulated ruminal environment. To achieve similar conditions of ruminal pH (acidity and alkalinity), buffered solutions were used (HCl, J.T. Baker®; NaOH, J.T. Baker®). A potentiometer (HANNA® instruments) was used to determine pH, which ranged from 5 to 7.5. Variations were sensed every 5 min. Incubation temperature was maintained between 34 and 44°C throughout the study. Fig 5 A and B show the results of linear regression of the data. Values of the experimental data (observed) were correlated with those estimated by the sensor. The results obtained show a significant relationship (p<0.05) in temperature (r=0.98) and in pH (r=0.97).

​

Once the system was calibrated in a cow rumen-like environment, the experiment was conducted inside the rumen of a cannulated cow. Since the diameter of the fistula was 11 cm, the food bolus design was adjusted to 10 cm in diameter, a length of 23 cm and a weight of 0.481 kg. The bolus was designed using SolidworksTM for later manufacturing. The electronics was designed to operate with power generated by a battery made up of three cells of 6 Amps to 3.7 volts Li-ion type (Fig 6A and B).

​

The device was introduced into the rumen of a previously fistulated cow for sensing during a period of 7 hours, taking 10 samples (data) every 30 minutes to obtain an estimate of the change that occurs when the system is placed in vivo. Since the measurements are in real time, we obtain a graph of the changes in temperature during the fermentation process and rumen pH over time. Changes in temperature and pH were observed from the time the animal started feeding, the period of voluntary consumption (Fig 6C).

Previous reports (Loholter et al., 2013), have evaluated the performance an experimental device in continuous measurement of pH and temperature, found important variations in the experimental data due to the phenomenon of drag, which is very common in instrumentation. However, the prototype was able to show moderate correlation between manual and continuous forms of measuring pH and temperature, similar to that found in our study. In other device designs (Kimura et al., 2012) have studied the relationship between pH and temperature in ruminal fluid to determine circadian changes in induced ruminal acidosis, evaluated both parameters in 10-minute intervals. They found significant modifications (p <0.05) of ruminal pH, which reached a low of 4 after 16 hours. In contrast, ruminal temperature had the opposite behavior, rising significantly (p<0.01) between 8 and 14 hours later, suggesting that active fermentation and the consequent production of volatile fatty acids (VFA) caused these modifications. The proposed equipment is capable of transmitting this condition during at least 24 hours of continuous use. In other work, a pH probe for continuous rumen pH measurement was used in a preliminary study in cattle during a 10-day period; probes were programmed to sample pH every 30 s, and sampling errors of 0.08 pH units, similar to our results (Enemark et al., 2003), were reported.  It has been commented (Eihvalde et al., 2016) that a prototype of these characteristics must be able to measure pH ranges between 6.8 and 5.4 (± 0.1) to be used in risk-prevention programs of ruminal acidosis. As the pH of rumen ingesta approaches 5.0, the amplitude and frequency of rumen contractions progressively diminish with eventual stasis (Kumar et al., 2007). The equipment under study meets this condition. The results obtained with this devices used to measure pH and rumen temperature in an experimental tropical cattle cow, will allow the development of new devices using micro and nanotechnologies, which allow access, permanence, temporary operation and send information from the rumen.

The prototype designed to transmit values of pH and ruminal temperature continuously meets the expectations of the interdisciplinary work group who developed it. It has the ability to operate in the desirable physiological ranges for clinical and productive experimentation, and correlation of observed data against estimates coincides with previous studies with similar prototypes, a condition also attributable to the capacity of data transmission for at least 24 hrs. Finally, as future work, it is considered to design and manufacture a prototype in integrated circuit technology which, given its dimensions, integrates all the modules of the electronic stage that senses the physiological parameters as well as transmits the data wirelessly in real time, allowing non-invasive testing.