Project Team


Students

Chang Bickel
Electro-Mechanical Engineering Technology
Penn State Altoona






Faculty Mentors

Sohail Anwar
Penn State Altoona
Engineering


Hussein Abdeltawab
Penn State Behrend
School of Engineering


Constantino Lagoa
Penn State University Park
School of Electrical Engineering and Computer Science






Project








Project Video




video player icon




Project Abstract


Chang Zhe Bickel

Major: Electro-Mechanical Engineering Technology
Penn State Altoona

Anticipated Graduation Date: Spring 2021

Faculty Advisor: Sohail Anwar (Altoona)
Hussein Abdeltaweb (Behernd)
Constantino Lagoa (University Park)

Project Title: Improving Object Detection Robot Performance

Detection Robots are used for many applications, such as the floor cleaning robot that avoids obstacles in its path. The output of the robot sensor is considered a Gaussian random variable in the presence of an object. The mean value of the output is one when the object is present and zero when the object is absent. The variant for the probability density function (PDF) is dependent on the quality of the sensor used for the robot, while low quality sensors usually have larger output variance. To improve the detection quality, this research proposes a decision algorithm that defines a specific threshold value. When the sensor output value is above this threshold, the robot will decide an object is present and vice versa. Utilizing high quality sensors is expensive which increases the robot cost. As an alternative solution, we are aiming to improve the performance of robot detection by the proposed algorithm without using high-quality sensors.

By combining the sum of PDF and weighted sum, calculating the weighted sum of multiple robots’ outputs will improve the detection system’s performance. The improvements were found by comparing the receiver operating characteristic (ROC) curves for different systems. The system will need to assign the best weights to each robot to ensure optimum performance. This system can be used in the future to make detection robots more accurate without the need to buy higher quality sensors.




Evaluate this Project


Use this form link to provide feedback to the presenters, and add your project evaluation for award(s) consideration.