ROBOT GRIPPER SENSOR OPERATION IN FRESH MANGO POSTHARVEST HANDLING FOR NON-DESTRUCTIVE FIRMNESS ASSESSMENT

Autores UPV
Año
CONGRESO ROBOT GRIPPER SENSOR OPERATION IN FRESH MANGO POSTHARVEST HANDLING FOR NON-DESTRUCTIVE FIRMNESS ASSESSMENT

Abstract

Mangoes quality is related with its firmness. Mangoes firmness is assess with a robot gripper while grasp the mangoes, which has accelerometers embedded on each of their fingers (A, B and C). 7 sets of 20 mangoes ¿Osteen¿, produced in Andalucia, were analyzed during 0, 2, 5, 7, 9, 12 and 14 days. Magoes were analyzed by sets first with the gripper and further in the laboratory; external and internal color, soluble solid content, total acidity and firmness by a cylindrical compression test. The gripper provides 24 variables but 12 were dismissed due to their low repeatability. With a factorial analysis, 6 factors explained 84% of the variability. The first is related to soluble solid content, destructive firmness, internal color and two variables extracted from the A and B accelerometers, the second factor to A and B accelerometers variables, the third factor to acidity, pH and maturity index, the fourth factor is related to external color, the fifth factor to the variables extracted from the C accelerometer and the sixth factor to the size variables. Results showed how the C accelerometer is not working as the A and B accelerometers. A partial least square regression model was carried out to explain destructive firmness with the variables extracted from the accelerometers A and B. From the 140 tested mangos a sample of 110 was randomly chosen as calibration set. A leave-one-out cross-validation was used with the calibration set. The model had a coefficient of determination (R2) of 0.73, the validation had 0.72 and the prediction had 0.80, with an average absolute error of 2.3 N/mm. The results showed the possibility of using A and B accelerometers from the robot gripper to non-destructively assess mango firmness during postharvest handling.