05-16-2023, 10:55 AM
Hello everyone,
I am currently involved in a project that focuses on developing a robotic grocery bin picking application. Our setup consists of a KUKA KR6 Agilus robotic arm controlled by a KRC4 Compact controller. To communicate with the controller, we utilize KukaVarProxy (KVP), and the KRL program running on the robot is the same used by RoboDK to send commands to the Kuka controllers.
Recently, we encountered an issue where the KRC would experience delays of several seconds in receiving commands. Upon investigation, we discovered that numerous Windows services were running on the KRC, which resulted in high CPU usage and subsequent communication delays. By disabling these services, we managed to eliminate the lengthy delays; however, the response times still lacked consistency (likely due to the nature of TCP/IP). On occasions, we observed delays of 120ms or even 300ms, leading to collisions.
During my research, I came across a new driver called c3bridge/Kuka Bridge, which is utilized by RoboDK. This driver is said to offer various advantages, and I am interested in understanding whether it provides any improvements in terms of the robot's response times to commands.
I stumbled upon a paper that confirms the stochastic nature of KVP. I would like to know if the new driver offers any enhancements in this regard and, if possible, quantify the extent of improvement it provides.
I am currently involved in a project that focuses on developing a robotic grocery bin picking application. Our setup consists of a KUKA KR6 Agilus robotic arm controlled by a KRC4 Compact controller. To communicate with the controller, we utilize KukaVarProxy (KVP), and the KRL program running on the robot is the same used by RoboDK to send commands to the Kuka controllers.
Recently, we encountered an issue where the KRC would experience delays of several seconds in receiving commands. Upon investigation, we discovered that numerous Windows services were running on the KRC, which resulted in high CPU usage and subsequent communication delays. By disabling these services, we managed to eliminate the lengthy delays; however, the response times still lacked consistency (likely due to the nature of TCP/IP). On occasions, we observed delays of 120ms or even 300ms, leading to collisions.
During my research, I came across a new driver called c3bridge/Kuka Bridge, which is utilized by RoboDK. This driver is said to offer various advantages, and I am interested in understanding whether it provides any improvements in terms of the robot's response times to commands.
I stumbled upon a paper that confirms the stochastic nature of KVP. I would like to know if the new driver offers any enhancements in this regard and, if possible, quantify the extent of improvement it provides.
