Umeå University's logo

This contribution presents the experience of a way to introduce the basic topics of laser range finders and simulation of autonomous drive for high school students. The aim was for them to get familiar with laser technology for autonomous system by being physical active by playing the role as artificial intelligence unit or the vehicle drive controller. In the play setting the student drove push tables up and down a corridor. At the end of the corridor they switched role, and the one that was acting artificial intelligence unit became vehicle controller, and the vehicle controller the artificial intelligence unit. Most students did try both roles. A minority of the students wanted to try several times, and a few students did not want to participate but preferred to observe or contribute as obstacles. It seemed the student where interested in the tasks and investigated the system response. Most of them drove the push tables fully controlled down the corridor in a playful way with other students as moving and stationary obstacles. The students also added new driving commands such as reverse that was initially not listed. The students were open to collisions as haptic feedback. It looked like they had lot of fun in the exercise. What can be mentioned is that most students were active. They did not only act vehicle controller and artificial intelligence unit, and they also acted stationary and moving obstacles, even if it was not mentioned in the briefing. They also did not need any previous knowledge as programming skills, mathematics, or geometry in order to participate. So, this exercise can easily be extended and improved by having a debrief discussion about the experience, and then introduce mathematics that can be used to solve the task in real life, such as trigonometry and geometry.

This paper presents a robotic course that was revised with the ICE (Ideas-, Connections-, Extensions levels) approach of teaching in mind. It includes practical labs where the students have to derive and implement mathematical models and verify those on a manipulator arm mounted on a small mobile robot called KUKA youBot. To support hands-on experience Simulink blocks have been developed to enable sensor readings and control of the youBot. The students implement their models using Simulink and run them to control the youBot manipulator. The labs have been used in several course settings with some modifications. The course also includes a small project where the students use and go beyond the results achieved in the labs, it relates to the Extensions level in ICE. Even though the approach is under development it has been tested and successfully used in a course.

We describe a project within the Design-Build-Test course where a student group, based on research, implemented help functions on a power wheelchair. The Design-Build-Test course at Umeå University comprises both an industrial relevant student projects and non-technical exercises like project management, teamwork (team dynamics) and communication. The goal is to create a learning environment where students from different study program work together in projects, resembling the conditions for projects in the industry. We believe that this approach will promote valuable skills in the field of product and system development which are important for the students’ future role as engineers.

We have developed a teleoperated quad-bike to be able to test a rollover prevention system. It was equipped with onboard computers, power steering, throttle control, emergency power cutoff, an IMU module, pneumatic actuators, and mechanical retractable outriggers. An onboard computer system tracks the vehicle orientation and velocity in real time. When the vehicle is predicted to overturn the system activates functions to prevent a rollover. The functions are: - activation of retractable outriggers, - decrease the throttle for motor brake, - activation of rear wheel brakes. The system has no active steering. The system was field tested in winter conditions in mars 2010, and was able to prevent rollover situations sideways and backwards.

This paper presents statistics on registrations from laser range finders in snowfall. The sensors are standard laser range finders in robotics, the LMS200 and the URG-04LX. Three different working cases were identified for the pulsed laser range finder. 1) Normal operation with background objects present within the range of the sensor. 2) Close range objects where ranges to objects are shorter than the pulse length. 3) Free-space in the background. The findings are summarized as: •  Two laser range finders have been used, one that sends out a pulsed wide beam and one with a modulated narrow laser beam. The narrow beam laser has better penetration between the snowflakes. •  Median filtering shows a substantial reduction in snowflake detects. •  The gamma distribution describes fairly well the range distribution of detected snowflakes. •  In an intense snowfall where about 24% of the ranges detected snowflakes. •  A time-polar median filter showed good results in suppressing snowflakes in range data.

In this paper we propose a tongue operated electric wheelchair system to aid persons with severe disabilities. Real-time tongue gestures are detected and estimated from a video camera pointing to the face of the user. The tongue gestures are used to control the wheelchair. Our system is a non-contact vision system, making it much more convenient to use. The user is only required to move his/her tongue to command the wheelchair. To make the system easy to drive, the system is also equipped with a laser scanner for obstacle avoidance. We also mount a 2D array of vibrators on the chair to provide the user with the information of the response from his/her tongue movement and the status of the system. We are carrying out user tests to measure the usability of the system.

This paper presents a method for obstacle avoidance and path-finding amid circular objects. The input data are circles and the output is a sequence of circles. The output circles represent a possible path, to a target, for a holonomic mobile robot. The method uses a solution to the Apollonius Tangency problem to find the maximum spanning circles amid the input circles. The radii of the circles can be set by desired clearance to nearby obstacles, from sensor parameters, or model parameters from extracted features. The method is intuitive and rather easily implemented and suits well for mobile robots, especially mobile robots with circular shape. It can be implemented with a depth-first approach where the target bearing angle is used as criteria in a divide and conquer step. The method was tested on sensor data registered by a laser range finder.

This paper presents how expanding circle sectors are used for finding free-space in front of a vehicle equipped with a laser range finder. The circle sectors give support for detection of oncoming cars since they will cause changes in the circle sector parameters. The velocities and headings of detected oncoming cars are estimated from clustered range data that are time stamped.