Automated Estrus Detection Using Activity Monitors and Pedometers

1. Introduction

• Increased milk production (higher estrogen metabolism)
• Larger herd sizes reducing individual cow attention
• Reduced time spent in free movement (confined housing)
• Increased incidence of silent estrus (ovulation without detectable behavioral signs)

2. The Estrous Cycle and Hormonal Basis of Activity Changes

2.1 Overview of the Bovine Estrous Cycle

2.2 Hormonal Control and Activity Increase

• Restlessness and increased locomotion — the cow walks more, often 2–4 times the normal step count
• Seeking behavior — the cow approaches and sniffs other cows
• Standing to be mounted (the primary, most reliable sign)
• Mucoid vaginal discharge
• Redness and swelling of the vulva

3. Traditional Methods of Estrus Detection

3.1 Visual Observation

• Observation 2–3 times daily (morning, noon, evening), each session ≥20 minutes
• Relies on identifying the primary sign: standing to be mounted (standing heat)
• Secondary signs: chin resting, mucoid discharge, vulval swelling, restlessness
• Detection efficiency: ~50–60% when done three times daily; far lower with once-daily observation

3.2 Heat Detection Aids

3.3 Hormonal and Physiological Indicators

• Progesterone assays (milk/blood): confirm non-pregnant, cycling cows
• Cervicovaginal mucus examination: clear, watery mucus increases at estrus
• Ultrasound: tracks follicular development, confirms ovulation

4. Automated Estrus Detection Using Pedometers

4.1 Principle of the Pedometer

4.2 Types of Pedometers

┌─────────────────────────────────────────────────────────────┐
│                   PEDOMETER TYPES                           │
├──────────────────────────┬──────────────────────────────────┤
│  Mechanical Pedometer    │  Electronic/Digital Pedometer    │
├──────────────────────────┼──────────────────────────────────┤
│ Spring/ball mechanism    │ Reed switch or piezoelectric      │
│ Simple step count        │ sensor counts steps electronically│
│ Manual reading required  │ Automatic data download via       │
│ Low cost                 │ transponders at milking parlour   │
│ Less accurate            │ Higher accuracy and sensitivity   │
│ Examples: basic leg      │ Examples: SCR Heatime® (leg),     │
│ pedometers               │ Afimilk AfiFarm®                  │
└──────────────────────────┴──────────────────────────────────┘

4.3 Placement and Operation

1. Cow walks → pedometer counts steps
2. Steps stored in hourly blocks
3. At each milking, the cow walks past a transponder/reader at the parlor entry
4. Data is transmitted wirelessly (radio frequency or Bluetooth)
5. Computer software compares current step count to the cow's rolling 7-day baseline
6. If activity ratio exceeds threshold (typically ≥200% of baseline), an estrus alert is generated on the farm management screen

                        PEDOMETER-BASED SYSTEM
                        
    [COW]                   [TRANSPONDER AT PARLOR]         [COMPUTER]
      │                              │                          │
  ┌───┴───┐                     ┌────┴────┐               ┌────┴────┐
  │Leg    │  ←step count→      │ Reader/ │  ←wireless→   │Software │
  │pedome-│                    │ Antenna │               │Database │
  │ter    │                    └────┬────┘               │Estrus   │
  └───────┘                         │                    │Alerts   │
                               Data download              └────┬────┘
                               at each milking                 │
                                                        ┌──────┴──────┐
                                                        │ Farm Manager│
                                                        │ Phone/Screen│
                                                        └─────────────┘

4.4 Interpretation of Pedometer Data

• The system calculates a ratio: Current activity ÷ Average baseline activity
• An activity ratio ≥ 2.0 (200% of baseline) over a 2-hour window is the common alert threshold
• Alerts are typically generated 12–24 hours before peak estrus, allowing timely AI

Step count
   ↑
   │           ██████
   │          ████████         ← ESTRUS ALERT THRESHOLD
   │   ░░░░ ████████████ ░░░░
   │░░░░░░░░░░░░░░░░░░░░░░░░░░
   └──────────────────────────→ Time (days)
     Day 19  Day 20  Day 21  Day 1
              ↑
          ESTRUS ONSET

4.5 Performance of Pedometers

5. Automated Estrus Detection Using Activity Monitors (Accelerometers)

5.1 What is an Accelerometer?

5.2 Mechanism of Action

         Z-axis (vertical — up/down)
              ↑
              │
              │──────→  Y-axis (lateral — side to side)
             /
            /
           ↙
      X-axis (longitudinal — forward/back)

1. Sampled at high frequency (e.g., 32–64 Hz)
2. Filtered using digital signal processing algorithms
3. Classified into behavioral states (walking, lying, eating, running/mounting)
4. Compared to the rolling baseline for that individual cow
5. Alert generated when the activity score deviates significantly

5.3 Ear-Tag Activity Monitors

• 3-axis head movement (proxy for feeding, rumination, and mounting-related head bob)
• Overall agitation scores

• Sensitivity: 82.9%
• Specificity: 98.3%
• Positive predictive value: 82.9%

5.4 Neck-Collar Activity Monitors

• Activity — via accelerometer
• Rumination — via microphone or jaw movement sensor (rumination time drops ~40% during estrus)
• Feeding time — feeding duration decreases at estrus

6. Integration with Farm Management Software

1. Stores individual cow data — rolling 7-day activity baselines
2. Generates ranked estrus lists — cows ranked by confidence of alert
3. Sends SMS/push notifications to farm managers
4. Integrates with milking robots — automatic sorting of cows flagged for estrus
5. Tracks conception results — linkage of AI date to pregnancy confirmation

6.1 System Architecture (Diagrammatic)

┌──────────────────────────────────────────────────────────┐
│               AUTOMATED ESTRUS DETECTION SYSTEM          │
│                                                          │
│  ┌──────────┐    ┌───────────────┐    ┌──────────────┐  │
│  │ Wearable │───▶│ Transponder / │───▶│    Central   │  │
│  │  Sensor  │    │   Antenna     │    │   Database   │  │
│  │(neck/leg/│    │ (at parlor,   │    │   Server     │  │
│  │  ear)    │    │  water trough │    │              │  │
│  └──────────┘    │  or gateway)  │    └──────┬───────┘  │
│                  └───────────────┘           │          │
│                                              ▼          │
│                                    ┌──────────────────┐ │
│                                    │  Farm Software   │ │
│                                    │  Dashboard       │ │
│                                    │ • Estrus alerts  │ │
│                                    │ • AI schedules   │ │
│                                    │ • Health flags   │ │
│                                    └────────┬─────────┘ │
│                                             │           │
│                                    ┌────────▼─────────┐ │
│                                    │  Farmer / Vet    │ │
│                                    │  Phone / Tablet  │ │
│                                    └──────────────────┘ │
└──────────────────────────────────────────────────────────┘

7. Commercially Available Systems

8. Comparative Evaluation: Pedometer vs. Accelerometer

9. Factors Affecting Performance of Automated Systems

9.1 Animal Factors

• Lameness: Lame cows show reduced baseline activity, making estrus-related increases harder to detect with leg pedometers
• Body condition score: Over-conditioned cows may show less activity during estrus
• Parity: Heifers show more intense estrus behavior than multiparous cows
• Postpartum period: Activity detection may be less reliable in early lactation (<60 days in milk)
• Silent estrus: Even the best automated systems miss some silent estruses

9.2 Environmental Factors

• Floor surface: Wet, slippery, or slatted floors reduce locomotion even during estrus, reducing system sensitivity
• Stocking density: Overcrowding restricts movement, dampening activity peaks
• Heat stress: High ambient temperature (>27°C THI) reduces estrus expression and duration
• Housing type: Loose housing with adequate space improves estrus expression and detection accuracy

9.3 System Factors

• Threshold setting: Setting the alert threshold too low → many false positives; too high → missed estruses
• Calibration period: Systems need 5–7 days of baseline data per cow before reliable alerts
• Sensor maintenance: Batteries, firmware updates, and sensor integrity must be regularly checked

10. Role in Improving Reproductive Efficiency

10.1 Benefits to Farm Management

• Improved estrus detection rate (EDR): Automated systems can achieve EDR of 75–95% compared to 40–60% by visual observation alone
• Optimal timing of AI: Activity data predicts the onset and peak of estrus; most systems recommend inseminating 6–18 hours after the alert
• Reduced inter-calving interval: Earlier and more accurate detection reduces days open and shortens calving intervals toward the 365-day target
• Labor savings: Eliminates the need for multiple daily observations; one farm worker can monitor a large herd remotely
• Economic benefit: A net gain of approximately $11.97 per heifer was found when switching from timed-AI protocols to activity monitor-based programs (Macmillan et al., 2021, PMID: 34555714)

10.2 Combined Use with Other Methods

• Tail painting (visual confirmation of mounting events — improves true estrus confirmation)
• Progesterone monitoring (rules out cystic ovaries; confirms cyclic status)
• Ultrasound (confirms dominant follicle and allows timely AI)
• Timed-AI programs (for cows with repeat missed alerts)

11. Limitations and Challenges

1. Initial investment cost: Hardware, software, and installation can be expensive, particularly for small herds
2. Technical complexity: Requires training of farm staff and reliable internet/network infrastructure
3. False alerts: Activity increases from other causes (social competition, regrouping, health events) can generate false estrus alerts
4. Battery and maintenance: Regular maintenance is needed; sensor failure leads to missed detections
5. Not suitable for all breeds/systems: Systems validated primarily in Holstein dairy cows; performance varies in beef breeds and extensively managed systems
6. Silent estrus still missed: No automated system achieves 100% sensitivity; silent estruses continue to be a challenge

12. Recent Advances

• Acoustic (vocalisation) detection: Wang et al. (2023, PMID: 37150135) developed a dual-channel acoustic ear tag that uses machine learning to identify characteristic vocalizations of cows in estrus, achieving high accuracy
• Rumen boluses (SmaXtec): Internal sensor boluses monitor core body temperature (which rises ~0.3°C during estrus) combined with movement data for highly specific detection
• Machine learning and AI: Deep learning algorithms applied to raw accelerometer data streams show promise for detecting not just estrus but also sub-clinical health events (ketosis, mastitis onset) from the same sensor
• Infrared thermography (IRT): Non-contact temperature measurement of the vulva and nasal planum to detect the temperature increase associated with estrus
• Image processing / computer vision: Camera systems with pose-estimation algorithms to identify mounting behavior in groups without any physical sensor on the animal

13. Conclusion

References

1. Rajput AS, Mishra B, Rajawat D, Bhakat M. (2024). Early prediction of oestrus for herd fertility management in cattle and buffaloes – a review. Reproduction in Domestic Animals, 59. PMID: 38798195
2. Merkelytė I, Šiukšcius A, Nainienė R. (2025). The Role of Sensor Technologies in Estrus Detection in Beef Cattle: A Review of Current Applications. Animals (Basel), 15(15):2313. PMID: 40805102
3. Macmillan K, Boyda A, Gobikrushanth M, Plastow G, Colazo MG. (2021). Economic comparison of an ear tag automated activity monitor for estrus detection with timed-AI in Holstein heifers. Theriogenology. PMID: 34555714
4. Schweinzer V, Gusterer E, Kanz P, et al. (2019). Evaluation of an ear-attached accelerometer for detecting estrus events in indoor housed dairy cows. Theriogenology, 130:19–25. PMID: 30856411
5. Schilkowsky EM, Granados GE, Sitko EM, et al. (2021). Evaluation and characterization of estrus alerts and behavioral parameters generated by an ear-attached accelerometer-based system for automated detection of estrus. Journal of Dairy Science, 104(5):5802–5817. PMID: 33685699
6. Madureira AML, Burnett TA, Borchardt S, et al. (2021). Plasma concentrations of progesterone in the preceding estrous cycle are associated with the intensity of estrus and fertility of Holstein cows. PLoS One. PMID: 34370740
7. Hubner AM, Canisso IF, Peixoto PM, et al. (2022). Effect of GnRH administered at AI for cows detected in estrus by conventional or automated activity-monitoring systems. Journal of Dairy Science, 105(1). PMID: 34756436
8. Marques JCS, Burnett TA, Denis-Robichaud J, et al. (2025). Tail chalk improved detection of true estrus alerts from an automated activity monitor system. Theriogenology. PMID: 40752313
9. Álvarez J, Acción A, López E, et al. (2025). Estrus Detection and Optimal Insemination Timing in Holstein Cattle Using a Neck-Mounted Accelerometer Sensor System. Sensors (Basel). PMID: 40942675
10. Wang J, Chen H, Wang J, et al. (2023). Identification of oestrus cows based on vocalisation characteristics and machine learning technique using a dual-channel-equipped acoustic tag. Animal, 17(6). PMID: 37150135