Sow longevity is a key component to breeding herd profitability. Lameness is a major reason for culling and therefore can negatively impact sow longevity. Lameness assessment in group housing systems requires the labor of skilled employees on a regular basis. The microcomputer-based embedded force plate system (hereafter referred as force plate) has been shown to identify lameness in laboratory-based settings. The goal of this research was to diagnose sow lameness using the force plate and pattern recognition methods in a field setting before lameness could be visually identified by a producer or by the herd veterinarian. In order to accomplish this, weight distributions measurements, produced by lame and sound sows using the force plate are needed; and then, relationships between leg weight distribution and different degrees of lameness can be determined. From this a lameness decision tree was developed to distinguish between sound and lame sows. To accomplish this the force plate was installed under one of two electronic sow feeders (ESF) in a group sow housing system with 120 multiparous sows for 21 days at the University of Pennsylvania – Swine Teaching and Research Center, Kenneth Square, PA. The force plate consists of four quadrants with non-slip coating (front right, front left, rear right, and rear left) which measures the force or weight that a sow applies per second to each quadrant (foot). Total dimensions were 1524 L x 565 W x 506 mm H, with 6.4-mm thick aluminum plating. Sow’s force measurements were recorded every second once pressure was applied to all four quadrants on the force plate while the sow consumed her daily feed allotment. If no pressure is applied to one quadrant or the sum of the pressure applied to two adjacent quadrants is less than 10 lbs., the value is deleted.

Lameness was also visually scored on a weekly basis using a four point assessment scale; where 0 = normal and 3 = severely lame. Sows were classified as non-lame (score ≤ 1) or lame (score ≥ 2). Based on the visual weekly assessments and the force plate readings, a lameness detection decision tree was created. This decision tree was consistent with visual assessments 95.98% of the time. Comparing the two lameness detection techniques, the force plate identified lameness on average 4.62 days sooner than the visual assessment. Through the trial, the force plate only required minor maintenance including calibration of quadrants and tightening of screws. One quadrants plate needed to be replaced as the non-slip surface on it had started to wear off. Results suggest that the force plate can detect lameness in a commercial setting earlier than weekly visual assessments while requiring minimal maintenance.

Key Findings:
• The force plate can be successfully installed in a commercial setting under an ESF with some mechanical maintenance and modifications including routinely calibrating the load cells to be sure measurements are accurate.
• A decision tree created based on the information gathered from the force plate can accurately detect lameness earlier than a weekly visual assessment.
• If producers using an ESF system can detect lameness earlier, then management practices for lame sows could be modified to trigger the gait on the ESF to send sows classified as lame into an isle or alternative pen where the stockman would easily identify them and provide a course of action for each sow.
• Further modifications are required to allow the force plate as currently designed to be commercially viable to collect data form a wide range of pigs (from gilts to mature boars).
• Future prototypes could include shorter front plates and a taller center bar to correct these problems.