Page 38

EETE APRIL 2013

ELECTRONICS IN ROBOTICS Robotics makes inroads into education By Julien Happich the emea freescale cup challenge that took place late last month at ESIE Paris’ university campus – www.esiee.fr - gathered 83 students, from 27 teams representing their respective universities from 10 European countries. The EMEA event itself was only one of many qualifying rounds within the global challenge organized by Freescale Semiconductor and involving university students from Brazil, China, EMEA, India, Japan, Korea, Malaysia, Mexico, Russia and the USA. The company provided basic car kits, either based on its Qorivva MPC5604B 32-bit automotive controller or its Kinetis ARM Cortex-M4 32-bit industrial MCU, and a set of limiting rules on hardware. With these basics and leveraging their programming skills, the students had to build the fastest intelligent robot car capable of following a sinuous track autonomously, without any remote control or guidance except that provided by its own on-board sensors. Spread over 180m2, the racetrack included speed bumps, intersections, hills and chicane curves, all featuring a centre line for optical detection. To help students with their designs, comprehensive online resources were also provided at www. freescale.com/freescalecup By distributing thousands of base kits to participating universities throughout the world (over 7000 for China alone), Freescale’s assumed goal is to encourage undergraduates to dip into robotics. The competition’s rules, which exclude the use of any auxiliary processor or other programmable device (apart from the company’s 32-bit MCUs), put the emphasis on optical sensing for the navigation, with a maximum of 9 sensors authorised on-board (16 for the global challenge). That means the participants must focus largely on sensor interfacing and signal processing, combining their newly acquired skills with clever embedded programming in order to succeed. By nurturing these young innovators Lined up for competition, some of the cars sport custom boards, others run the full Freescale Tower development system. The Freescale Cup Kit 1/18 scale model car chassis. with fresh hardware and software design props, Freescale also hopes to get a head start over competition in what will be tomorrow’s embedded systems. “We have been running these Freescale Cup Challenges for a number of years now (since 2003), and automotive system manufacturers keep a close eye on the winners. In fact, Continental often end-up recruiting some of the best contestants”, told me Laurent Massicot, Media Relations Manager EMEA for Freescale Semiconductor. ”If we are looking for new talents to fulfil our own engineering positions, the design entries and ideas we receive give Freescale an interesting pool of potential candidates”, he added. The Freescale Cup Kit includes a 1/18 scale model car chassis equipped with one 7.2V DC motor for each rear wheel, a servo motor for the steering, the company’s MCU interface and motor controller board, a CMOS camera and/or IR sensors, and a 128-pixel line scan camera with a focusable imaging lens. Power-wise, the rechargeable NiCd or NiMH battery provided by the participants is limited to a maximum capacity of 3000mAh. There are also limits set on any additional capacitors and the overall dimensions of the fully assembled car. Depending on how heavily the algorithms treat sensor data, the MCU may draw too much power at the expense of propulsion, or simply take too long to process the data, slowing things down. In that case, the car may not even cover the distance of a full lap. The specifications of the main track modules (straight lines, curves and hills) had been provided before hand so the universities could build some experimental racetracks, but the final design was only disclosed during the event. Prior to the final race, students had a day and a half to test and fine tune their vehicles. They also had to demonstrate their car’s ability to avoid obstacles, quickly accelerate and brake within a given distance. Another challenge was to use Freescale’s newest “eCompass” sensor technology to drive their model cars without a track and to come back to a pre-set point. All these challenges relate directly to some of the new features that can typically be found in modern driver assistance systems. For the final, two laps were recorded for each car, with only the best time taken into consideration. Going off-track, even for a second or partially (a tire overlapping an edge) was considered as a failure, and the referees were strict. Various racing strategies were adopted. Some teams performed the two laps independently, using mostly a control loop to evaluate track curvature based on the centre line’s deviation from one data acquisition cycle to the other (from the line scan camera), adjusting their steering and speed accordingly. Some others clearly adopted an exploration strategy, running their first lap very slowly to exploit as many sensors as possible 38 Electronic Engineering Times Europe April 2013 www.electronics-eetimes.com


EETE APRIL 2013
To see the actual publication please follow the link above