Learning Embedded And Real Time Systems Via Low Cost Mobile Robotics

Theimportanceandimpactof embeddedandreal-timecomputingsystemsontoday’ssocietyfar exceedthatof traditionalstand-alonecomputers;it is hardto think of new device or systemthat doesnot embedsomecombinationof silicon-basedintelligence,sensing,andcommunication. A paralleltrendis thegrowth of high-level, abstractdesignof thesesystemsandassociatedlanguagesandCAD tools,drivenby increasingprocessorspeed,time-to-market requirementsand thecomplexity of applications.However, many coursesin embeddedsystemsfocuson low-level issuessuchasaddressing, interruptsandinterfacing.In thispaper , wedescribeanew direction:a coursewith thegoalof motivatingstudentsto learntheabstractconceptsthatunderlythedesign of thesesystemsvia experimentsthatrequiretheinteractionof robotswith thephysicalworld. In this studio-formatcourse,mostconventionallearningtakesplaceoutsidetheclass,while small studentteamsdesign,build, andevaluateautonomousmobilerobotsin theclassroom/laboratory . To keepcostsdown, mobile robotsarecreatedusingLEGO partsandprogrammedin thehighlevel NQClanguageusingtheRobolabRCX microcontrollermodule.As thesemester proceeds, studentstackleanarrayof interrelatedproblemsthatmotivatethestudyof sensorsignalprocessing, control,scheduling,andresourcesharing.In a final project,thestudentstacklea distributed intelligenceproject in which an odometry-equippedrobot communicateswith a PC-basedprogramthattrackstherobot’s position.To encourageadoptionby otherelectricalengineeringand computerengineeringprograms,a detaileddescriptionof therequiredresourcesandtheir costis included.