Weßeler, Peter
Refine
Document Type
Language
- English (2)
Has Fulltext
- no (2)
Is part of the Bibliography
- no (2)
Keywords
- Algorithm learning (1)
- Author Keywords: Industrial robot, Automatic robot programming, SME, low volume, high variant, path planning, matching, machine vision (1)
- IEEE Keywords: Solid modeling, Path planning, Three-dimensional displays, Robot kinematics, Pipelines, Task analysis (1)
- INSPEC (Controlled Indexing): cameras, collision avoidance, industrial robots, mobile robots, small-to-medium enterprises (1)
- INSPEC (Non-Controlled Indexing): high variant manufacturing, industrial robots, complex robot tasks, medium-sized enterprises, camera-based path, planning overhead, fast program generation, collision-free path, specific robot task, SME (1)
- Language-independent programming (1)
- Tutoring system (1)
Teaching People to program is a crucial requirement for our society to deal with the complexity of 21st-century challenges. In many teaching systems, the student is required to use a particular programming language or development environment. This paper presents an intelligent tutoring system to support blended learning scenarios, where the students can choose their programming language and development environment. For that, the system provides an interface where the students request test data and submit results to unit test their algorithms. The submitted results are analyzed by a machine learning system that detects common errors and provides adaptive feedback to the student. With this system, we are focusing on teaching algorithms rather than specific programming language semantics. The technical evaluation tested with the implementation of Mean and Median algorithm shows that the system can distinguish between error cases with an error rate under 20%. A first survey, with a small group of students, shows that the system helps them detect common errors and arrive at a correct/valid solution. We are in the process of testing the system with a larger group of students for gathering statistically reliable data.
Camera based path planning for low quantity - high variant manufacturing with industrial robots
(2019)
The acquisition costs for industrial robots have been steadily decreasing in past years. Nevertheless, they still face significant drawbacks in the required effort for the preparation of complex robot tasks which causes these systems to be rarely present so far in small and medium-sized enterprises (SME) that focus mainly on small volume, high variant manufacturing. In this paper, we propose a camera-based path planning framework that allows the fast preparation and execution of robot tasks in dynamic environments which leads to less planning overhead, fast program generation and reduced cost and hence overcomes the major impediments for the usage of industrial robots for automation in SMEs with focus on low volume and high variant manufacturing. The framework resolves existing problems in different steps. The exact position and orientation of the workpiece are determined from a 3D environment model scanned by an optical sensor. The so retrieved information is used to plan a collision-free path that meets the boundary conditions of the specific robot task. Experiments show the potential and effectiveness of the the framework presented here by evaluating a case study.