Ground instability, such as roof and rib failures, can lead to damages to equipment, physical injuries of personnel, and some fatalities every year.Despite various advances in drilling and ground support techniques,ground instability remains to be a persistent safety issue.To improve underground mine safety and decrease roof fall incidents,an effective ground support design should be implemented.Optimization of ground support design is a function of rock mass properties.The ground related properties of interest,including characterizing joints, voids, cracks, and bed separations, are critical factors that control the roof and/or rib failures.Identifying distributions of these features,while performing routine drilling for rock bolt installation,allows for timely adjustment of ground support measures.This research focuses on locating of joints and classifying rock strengths by monitoring drilling parameters from an instrumented roof bolter during the operational cycle of drills.A series of full scale laboratory drilling tests were carried out on a roof bolter testing unit with various simulated-joint conditions, including joints with the aperture less than 3.175 mm, joints at different angles, changes of rock-strength on both sides of the joint.Additionally, detection programs based on updated pattern recognition algorithms were developed towards joint detection.This paper will briefly review testing procedures, data analysis, logic/algorithms, and summary of suggested improvements on joint detection programs based on the observations in drilling tests.