As the students of the 24 teams began to assemble at the Gymnasium Kirchheim bei München for the sixth qualification event of the NXP Cup, curiosity about each other’s 1/18 scale car models continued to build—with everyone comparing the modifications they made to their respective starter kits in order to master the challenges that lay ahead. The international flavor of the event was also very much evident, as teams arrived from Italy, Germany, the UK, North Africa, and Eastern Europe (Figure 1).
Figure 1: The participating teams in the NXP Cup. (Source: Author)
Luckily, on this particular day in late March, the weather was fairly cloudy; a great advantage as it meant that the compact vehicles would have fewer issues dealing with changing light conditions (something that proved to be problematic at the previous year’s event). This time there were no unexpected rays of sunshine creating awkward reflections or shadows on the black and white racetrack laid in the high school’s sports hall (Figure 2).
Figure 2: The numerous tracks that the vehicles from each team needed to negotiate in the NXP Cup. (Source: Author)
There was space for only four teams to qualify and join the total complement of 20 to compete on the 29th and 30th of April at the Fraunhofer IIS Institute in Erlangen, Germany—the venue for the final for the fifth year in a row. The incentive to succeed is very well defined, with NXP Semiconductors rewarding the team named EMEA Champion with a trip to Santa Clara, California to attend the company’s flagship NXP Connects event on the 12th and 13th of June.
The 2019 NXP Cup EMEA has seen over 150 student teams from high schools and universities across the region enroll in this unique competition for developing autonomous vehicles—with much of the focus placed on programming the small racecars for precision driving. This season NXP allowed participating students to expand the technologies deployed and respond to the requirements of the new optional trials—including the obstacle avoidance trial, the speed control zone, and the precision figure-eight track. Of course, the traditional and mandatory speed race is still an essential part of the challenge… always bringing a great deal of excitement.
The base kit for each car included a pair of DC brush motors for propulsion, a digital servo motor for steering purposes, a power management interface and NXP’s FRDM-KL25Z board (at the heart of which is a Kinetis L ARM® Cortex™-M0+ microcontroller) for the motor and guidance controls (Figure 3). The package also came with a simple CMOS 128-bit line scan camera to detect the signage and course markings on the black and white track. The onboard default 2500mAh, 7.2V Nickel-Cadmium (NiCd) battery offers the power capacity for vehicles to complete many laps. Mouser Electronics provided support, with the company acting as a key sponsor for supplying the kit, plus other useful technologies that students might want to add to their racecars—such as OSEPP’s ultrasonic sensors or higher quality CMOS cameras like the Pixy 2 units equipped with NXP’s LPC4330 microcontrollers. The teams relied on established programming and debugging platforms to prepare their vehicles for the different tasks encompassed within the competition—such as Mbed, Mathworks’ Matlab/Simulink or the MCUXpresso Integrated Development Environment (IDE).
Figure 3: The NXP Cup ‘pit lane’ showing all the competing racecars. (Source: Author)
Local team Hoibnbeschleuniger 2000, from the Munich University of Applied Sciences, decided to push the technology envelope by combining the image processing capabilities of NXP's ARM iMX6Q 1GB microprocessor with an OV5640 serial camera for machine vision, plus the efficiency of FRDM-KL25Z based motor control functions. A little anecdote here is that the team had to wait a few seconds for their car to boot up and load the Embedded Linux operating system before running it on the track. This differentiated it from the cars powered only via the standard MCU board, which were ready to run from the moment they were turned on. The team faced some issues getting the drivers to interface correctly with all the components in the car. A lack of development time meant they weren’t able to obtain the results they were hoping for. Nevertheless, they are fully prepared to continue working on it. They plan to move to a Yocto Project platform to have a better package manager and drivers for next season when they are looking to take another shot at the EMEA title.
One of the standout teams at the event came all the way from the Republic of Kosovo’s capital, Prishtina. Competing under the name BRKTeam, this collective consisted of students from the University of Prishtina who heard about the contest while visiting the NXP booth at the Electronica trade show last November and subsequently decided to get involved. They worked diligently on their car over the last few months, validating the technology requirements for each of the four principal challenges. The team predominantly concentrated on getting as much control as possible over every element of the vehicle by developing high-end algorithms for precision driving. One of the modifications implemented was to add an NXP H-Bridge to control the steering servo, rather than relying on the electronics of the servo itself. This provided them with extremely exact readings regarding the servo’s positioning and significantly improved vehicle handling. The most notable work was done in relation to the obstacle avoidance challenge, where the car had to steer clear of a 20cm x 20cm x 20cm white polystyrene cube while not exiting the track. To do so, they tried several sensor types, such as ultrasonic ones, but ended up going with a Slamtec A2 RPLIDAR interfaced with the Freedom board using the UART port. This provided accurate information about the vehicle’s surroundings. The LIDAR apparatus was designed to be removable, as it would not be needed for the other challenges and the speed race. In this way, they could reduce the risk of damage if an uncontrolled collision was to occur and make the vehicle lighter (and thereby quicker), using only the standard CMOS linescan camera for line detection work. For the EMEA finals, the team will focus on making the code more robust, especially the algorithm for the detection of the lines on the track.
The qualification round in Kirchheim was closely contested, with the BRKTeam taking top spot and DIT Deggendorf’s K-Team coming in second. Team Koala-Race from the University of Applied Sciences in Landshut, Germany made it to third place, followed by Team MIDI-Unimi from the Università degli Studi di Milano in Italy, who were in fourth position.
Do not forget to follow Mouser Electronics on social media (@MouserElecEU) to stay up-to-date with the next phase of the qualification process and the build up to the EMEA finals at the end of April.
Flavio Stiffan joined the semiconductor industry in 1989 and assumed several global roles focusing on global project management and business development. He holds an equivalency of a bachelor’s in computer science and a master certificate in marketing strategy, branding, business management from Cornell University.
In 2016, he founded Stiffan Consulting as freelance marketing coach and digital content creator focusing on knowledge transfer activities, market strategy development and brand positioning. He is based in the Munich area (Germany).