Paper deep dive
TS-MAMP: A Remanufactured Agricultural Robot Powered by Second-Life EV Components and NMS-Free On-Device Weed Detection
Weijie Shi, Zicheng Xu, Zhenbang Cheng, Haoran Xuan, Mingbo Duan, Gan Ge
Intelligence
Status: not_run | Model: - | Prompt: - | Confidence: 0%
Entities (0)
Relation Signals (0)
No relation signals yet.
Cypher Suggestions (0)
No Cypher suggestions yet.
Abstract
Abstract:Agriculture 4.0 robotic systems improve field efficiency yet remain too capital-intensive for the fragmented smallholdings that dominate global agriculture. Meanwhile, a growing number of retired low-speed electric-vehicle (LSEV) powertrains retain functional electromechanical value but are destructively recycled. This paper presents TS-MAMP (Telescopic-Sleeve Modular Agricultural Mobile Platform), a remanufactured robot built under 3R (reduce, reuse, recycle) circular-economy principles. Retired 48 V brushless-DC (BLDC) hub motors are paired via back-EMF matching, and lead-acid battery modules screened at 60%-80% state of health are actively balanced within a 100 mV inter-module voltage deviation. Together, these reused components reduce the powertrain-and-chassis BOM cost by approximately 60%, to below USD 450 (perception and weeding modules excluded). The truss chassis provides >=200 kg static load, continuously adjustable track width from 1200 mm to 2000 mm, and <=5-minute module changeover. An NMS-free (non-maximum-suppression-free) YOLOv10n detector with consistent dual-assignment training and negative-sample learning achieves 80.87% mean average precision (mAP)@0.5 (58.41% mAP@0.5:0.95) on the Wanxi Crop-Weed dataset, and is deployed via FP16 TensorRT on a Jetson Nano, confirming on-device inference feasibility. TS-MAMP demonstrates that retired EV components, under modest screening, can be re-engineered into affordable, AI-enabled agricultural robots--opening a remanufacturing pathway for the smallholder fields that commercial automation leaves unserved.
Tags
Links
- Source: https://arxiv.org/abs/2608.02270v1
- Canonical: https://arxiv.org/abs/2608.02270v1
Trouble viewing inline? Open PDF directly â
Full Text
21,456 characters extracted from source content.
Expand or collapse full text
TS-MAMP:ARemanufacturedAgriculturalRobot PoweredbySecond-LifeEVComponentsand NMS-FreeOn-DeviceWeedDetection WeijieShi Âč,â ,ZichengXu ÂČ,Âł,*,â ,ZhenbangCheng Âč ,HaoranXuan Âč ,MingboDuan Âč ,GanGe Âč 1 SchoolofMechanicalandAutomotiveEngineering,WestAnhuiUniversity,Lu'an,237000,China 2 SchoolofAutomation,NanjingUniversityofInformationScienceandTechnology,Nanjing,210000,China 3 SchoolofEngineering,WestlakeUniversity,Hangzhou,310030,China â Theseauthorscontributedequallytothiswork. * Correspondingauthors:202542000226@nuist.edu.cn AbstractâAgriculture4.0roboticsystemsimprovefield efficiencyyetremaintoocapital-intensiveforthefragmented smallholdingsthatdominateglobalagriculture.Meanwhile,a growingnumberofretiredlow-speedelectric-vehicle(LSEV) powertrainsretainfunctionalelectromechanicalvaluebutare destructivelyrecycled.ThispaperpresentsTS-MAMP (Telescopic-SleeveModularAgriculturalMobilePlatform),a remanufacturedrobotbuiltunder3R(reduce,reuse,recycle) circular-economyprinciples.Retired48Vbrushless-DC (BLDC)hubmotorsarepairedviaback-EMFmatching,and lead-acidbatterymodulesscreenedat60%â80%stateof healthareactivelybalancedwithina100mVinter-module voltagedeviation.Together,thesereusedcomponentsreduce thepowertrain-and-chassisBOMcostbyapproximately60%, tobelowUSD450(perceptionandweedingmodulesexcluded). Thetrusschassisprovidesâ„200kgstaticload,continuously adjustabletrackwidthfrom1200mmto2000m,and â€5-minutemodulechangeover.AnNMS-free (non-maximum-suppression-free)YOLOv10ndetectorwith consistentdual-assignmenttrainingandnegative-sample learningachieves80.87%meanaverageprecision(mAP)@0.5 (58.41%mAP@0.5:0.95)ontheWanxiCrop-Weeddataset, andisdeployedviaFP16TensorRTonaJetsonNano, confirmingon-deviceinferencefeasibility.TS-MAMP demonstratesthatretiredEVcomponents,undermodest screening,canbere-engineeredintoaffordable,AI-enabled agriculturalrobotsâopeningaremanufacturingpathwayfor thesmallholderfieldsthatcommercialautomationleaves unserved. Keywords-Agriculturalrobotics;EVcomponent remanufacturing;Circulareconomy;Edge-deployedweed detection;NMS-freedetection I.INTRODUCTION Manyhigh-throughputAgriculture4.0roboticsystems aredesignedforlarge,standardizedfarmlandandrequire substantialcapitalinvestment.TheNEXATgantry,withits ~1,100hpdrivetrainand200t/hharvestingmodules, representsamultimillion-poundinvestmentdesignedfor verylarge-scalecommercialfarms [1],[2] âillustratingthe starkgapbetweenhigh-endautomationandthesmall, fragmentedholdingstypicalofsmallholderagriculture [3],[4] . ModularrobotssuchasThorvaldIIhaveimproved configurabilityacrossfield,greenhouse,andspecialty-crop environments [5],[6] ,whilecommercialsystemslikeNaĂŻo Technologies'Tedhavedemonstratedtheprogressof autonomousmechanicalweedinginspecialtycrops [7] .Yet smallholderadoptionremainsconstrainedbycapitalcost, financialrisk,limitedcreditaccess,anduncertaintyabout returns [8]â[10] . Meanwhile,rapidtransportelectrificationisgeneratinga growingstreamofretiredEVcomponents.InChina,retired NEVpowerbatteriesareprojectedtoexceedonemillion tonnesannuallyby2030 [11] ,andcomprehensivereviews haveconfirmedtheviabilityofrepurposingtheseretired batteriesforsecond-lifeapplications [12] .Additionally, electrictractionmotorscontainvaluableassembliesthatare oftenlostwhenend-of-lifetreatmentreliesondestructive recycling [13]â[15] .Second-lifestudiesfurtherindicatethat batteriesretiredatapproximately70%â80%SOHretain usefulcapacityforlessdemandingstationaryorlow-speed applications [16]â[19] ,suggestinganopportunitytoconnect reuse-orientedremanufacturingwithlow-speedagricultural roboticplatforms. Inthiswork,wepresentTS-MAMP,aremanufactured agriculturalrobotbuiltfromsecond-lifeLSEVcomponents under3Rcircular-economyprinciples [20]â[23] .Unlike high-endgantryormodularplatformsthatrequirenew componentsandsubstantialcapital,TS-MAMPtargets fragmentedsmallholderterrainthroughatruss-style, dimensionallyreconfigurablechassispoweredbyretired BLDCmotorsandlead-acidbatterymodules,pairedwithan NMS-freeYOLOv10nperceptionsystem [24] thatremoves theneedforaseparateNMSpost-processingsteponedge hardware.Thecontributionsarethreefold:(1)asecond-life EVpowertrainqualificationandmodularchassisintegration methodologyreducingthepowertrain-and-chassisBOMcost byapproximately60%tobelowUSD450(perceptionand weedingmodulesexcluded);(2)prototype-levelmechanical verificationofâ„200kgstaticload,continuouslyadjustable trackwidthfrom1200mmto2000m,andâ€5min tool-modulechangeover;and(3)anNMS-freeYOLOv10n pipelineachieving80.87%mAP@0.5(58.41% mAP@0.5:0.95)ontheWanxiCrop-Weeddataset,deployed viaFP16TensorRTonaJetsonNanoforon-device inference.Thefollowingsectionsdescribethesystemdesign, experimentalsetup,fieldvalidation,andfuturework. I.SYSTEMDESIGN A.GreenRemanufacturingandEVComponentQualification TheTS-MAMPplatformwasdesignedandfabricated following3Rcircular-economyprinciples [20]â[23] .The remanufacturingworkflowprioritizeddirectreuseof functionalelectromechanicalassembliessalvagedfrom retiredlow-speedelectricvehicles(LSEVs),minimizing raw-materialextractionandmanufacturingenergy.Paired DJZ48-10C148V/350WBLDChubmotorswereselected asthetractionactuators.Salvagedunitswerematchedby back-EMFwaveformandinternalresistancepriorto installation,preservingkinematicsymmetrybetweentheleft andrightdriveunits. Forenergystorage,retired12Vgraphene-modified lead-acidbatterymoduleswerecollectedfromLSEV scrappingstationsandevaluatedwithreferencetothe capacityandsafetytestmethodsspecifiedinGB/T 32620.1-2016 [25] .Modulesretaining60%â80%ofrated capacitywereselectedandreassembledintoa48V (four-series)/20Ahtractionpack.The60%â80%retention windowwasestablishedthroughpreliminarydischarge characterizationof20retiredmodulesandreflectsan explicittrade-offbetweenusableresidualcapacityand expectedcyclelifeunderintermittent,low-current agriculturaldutycycles.Thisapproachredirectsretired batterymodulesfromuncontrolleddisposalintoacontrolled reusepathway,reducingtheriskofelectrolyteleakageand heavy-metalcontaminationassociatedwithimproper end-of-lifetreatment.Thequalifiedpowertrainconfiguration issummarizedinTableI. TABLEI.C ONFIGURATIONOFREMANUFACTUREDPOWERTRAINANDCHASSIS ComponentSource(RetiredEV) Specificationin Robot Energy Storage 12V20AhGraphene-modified Lead-acidBatteryĂ4 48VSystem(4 Series) Traction Motor 48V350WHubBrakeMotor 48V350WBLDC Actuator Transmission38-LinkChain&48T/16TGear ChainDrive (ReductionRatio3:1) ControlUnitBrushlessEVController Integrated Dual-ModePWM Driver ChassisNode FrontWheelHub(Model 21-44-20) DifferentialSteering System Thefullyassembledprototypewasbuiltusingretired components,anin-housefabricatedsteelframe,and off-the-shelfcontrolelectronics,withaprototype-leveldirect hardwareBOMcostbelowUSD450coveringthetraction powertrainandchassisonly(motors,batteries,frame, controller,batteryequalizer,andmechanicaldrivetrain;the perceptionstackâJetsonNanoandIMX219cameraâand thelaser-weedingmoduleexcluded),including transportation,screeninglabor,andfabricationoverhead, approximately60%lowerthananequivalent new-componentbaselinecomprisingcommerciallysourced 48VBLDCmotors,newlead-acidtractionbatteries,anda matchingframestructure. B.MechanicalArchitectureandDrivetrain Themechanicalplatformwasdesignedasatwin-parallel brackettrussframejoinedbyacentralsupportbase.The load-bearingstructurecombinesahigh-strength aluminum-alloyportalframewithQ235low-carbon-steel beams,providingacompromisebetweenstructuralrigidity, reducedframeweight,fabricability,andmaterialcost.The ratedstaticloadcapacityofâ„200kgwasverifiedthrough physicalprototypeloading.Thepowertrainadoptsachain transmissionratherthanabeltdrivetoeliminateslippage underhigh-dustandhumidfieldconditions.A3:1sprocket reduction(16-toothdrivingsprocket,48-toothdriven sprocket)mechanicallytriplestheoutputtorqueatthewheel hubs,creatingatorquesafetymarginthatcompensatesfor theexpectedelectromechanicalattenuationofsecond-life motors.Thismechanicalamplificationallowstheaged BLDCactuatorstooperatebelowtheirpeaktorquedemand duringnormaltraction.Theoverallmechanicalarchitecture andpowertrainlayoutareillustratedinFig.1. Figure1.Mechanicalarchitectureandpowertraincomponents. Toaccommodatevariablecroprowspacingandmodule reconfiguration,theframeintegratestwoindependent adjustmentmechanisms:(i)acontinuoustelescopic-sleeve systemenablingcontinuouslyadjustabletrackwidthfrom 1200mmto2000m,and(i)anindependent sliding-groovesystemprovidinglateralmodule-width adjustmentfrom300mmto600mmat±5mmpositioning accuracy.Bothmechanismsaresecuredby vibration-resistantlockingbolts,allowingcomplete tool-modulechangeoverwithin5minutes. C.MultifunctionalModulesandLaserWeeding TheTS-MAMPplatformutilizesstandardizedinterfaces andrapidconnectionmechanismstoenableconfigurationof multifunctionalmodules.Aprimaryconfigurationisthe laserweedingsub-module,aneco-friendlyalternativeto chemicalweeding [26],[27] .Thismoduleemploysa3-DOF Deltaparallelmanipulatorcomprisingthreeidenticallimbs actuatedby42mmhybridsteppermotorswith0.48N·m holdingtorque.Aparallelogramguidelinkageprovides positioningofafour-stagelasersystemwith4mmspot diameterand50Wopticaloutputpower.Thelaserassembly includesaprotectiveenclosuredesignedwithreferenceto theprotectivehousingrequirementsofGB/T7247.1-2024 [28] . Laseractivationispreciselycontrolledviathe Modbus-RTUcommunicationprotocolandintegratedwith safetyinterlockdevices.Whenthesystemdetectsthatthe platformmovesataspeedbelow0.1m/sorfailsto accuratelyidentifyworkingtargets,itactivatesthelaser outputsuppressionmechanismtopreventlocalizedthermal accumulationduringprolongeddwellandpotential secondaryenvironmentalhazards.Themultifunctional moduleconfigurationisshowninFig.2. (a)Moduleschematic(b)Laserweedingsubmodulestructure Figure2.Multifunctionalmoduleconfiguration. D.ElectronicControlandDriveSystem Theelectroniccontrolsystemintegratespower management,drivecontrol,andonboardperception, enablingcost-effectiverepurposingofretiredEV componentsforoperator-supervisedfieldoperation.The powertrainusessalvagedDJZ48-10C1BLDChubmotors (48V,350W),whichareequippedwithintegrated mechanicalexpansionbrakes,asillustratedinFig.3. (a)DJZ48-10C1BLDChubbrakemotor(b)Workflowschematic Figure3.Electroniccontrolanddrivemodule. Adual-channelmixingcontrollerservesasthecentral electronicinterfacebetweenthewirelesscommandsystem andthedriveactuators.Ittranslates2.4GHzPWMsignals receivedfromaMicrozoneMC6REradioreceiverinto calibratedanalogset-pointsfortwo500WBLDCmotor speedcontrollerswhilemaintainingalow-latencymanual overrideviathewirelesslinkforoperationalsafetyduring systemdebuggingandfieldmaneuvers. Toensureoperationalsafetyandalleviatevoltage deviationinretiredbatteries,thesystemisequippedwithan ANGUIKBX104Sactivebatteryequalizer.Controlledbyan MCU,thisdeviceusesadynamicenergytransferstrategyto maintaininter-modulevoltagedifferentialbelow100mV duringbothchargeanddischargephases.Forthermal management,modulesarephysicallyseparatedby15m gapswithinaventilatedchassistofacilitatepassive air-coolingdissipation. I.DATASETANDEXPERIMENTALSETUP A.WanxiCrop-WeedDataset Toaddressthedomaingapbetweenpubliccrop-weed datasetsandlocalfieldconditions,weconstructedtheWanxi Crop-WeedDatasetinLu'an,AnhuiProvince.Imageswere acquiredusingtheplatform'sonboardSonyIMX219CMOS cameraat1920Ă1080resolutionundernaturalillumination of20,000â100,000lux.Alldatawerecapturedduringthe 3â5leafseedlingstageofpakchoi,whenmorphological similaritybetweencropandassociatedweedsismost pronounced.Theinitialdatasetcomprised675images retainedatnativeresolution;formodeltraining,imageswere resizedto640Ă640pixelsusingletterboxpaddingto preserveaspectratio.The675imageswererandomlysplit intotraining(65%),validation(25%),andtest(10%)sets. Annotationfollowedaplant-wisebounding-boxprotocol coveringfiveclasses:pakchoi(Brassicarapasubsp. chinensis)andfourassociatedweeds:Echinochloacrus-galli, Eleusineindica,Digitariasanguinalis,andPortulaca oleracea. B.NMS-FreePerceptionPipelineandOptimization WeadoptedYOLOv10nwithconsistentdualassignment (CDA),whichsupportsNMS-freeinferencethroughthe YOLOv10dual-assignmentdesign [24] .Anauxiliary one-to-manyheadsuppliesrichsupervisorysignalsfor convergence,whileaprimaryone-to-oneheadassigns exactlyonepredictionperground-truthobjectusingastrict matchingmetric.Duringinference,onlytheoptimized one-to-oneheadisretained,removingtheneedforaseparate NMSpost-processingstepandenablingdeterministicedge deployment.TrainingwasperformedonanNVIDIARTX 3090(24GB)workstation;inferencewasdeployedonan NVIDIAJetsonNanoviaTensorRTFP16precision,a deploymentrouteshowntobefeasibleonsuchlow-power edgeplatformsbyrecentbenchmarkstudies [29] . Toaddressslowconvergenceandhighfalse-positive ratesincomplexseedling-stagescenarios,atwo-stage optimizationstrategywasapplied.StageIexpandedthe 439-imagetrainingpartitionto6,020imagesviaoffline dynamicmosaicandphotometricaugmentation(random horizontalflipping,brightnessscalingof±0.2,andhue shiftingof±0.1).StageIIintegratedahybridspatial/channel attentionmechanism(SAM/CAM)andintroduced500 target-freenegativebackgroundimages,bringingthetotal training-timesamplecountto6,520witha7.67% negative-sampleratio.Allaugmentationsandnegative sampleswereappliedexclusivelytothetrainingpartition. TableIIreportstheablationresults.TheNMS-free YOLOv10nbaselineachieved62.15%mAP@0.5(45.32% mAP@0.5:0.95). TABLEII.QUANTITATIVEABLATIONOFTHENMS-FREEYOLOV10NPERCEPTIONPIPELINE. OptimizationModulesPrecision(%)Recall(%)mAP@0.5(%)mAP@0.5:0.95(%)ÎmAP@0.5(p) Baseline72.2867.3462.1545.32â +DataAugmentation77.0371.2969.4649.37+7.31 +SpatialAttention79.6574.1374.2252.81+4.76 +ChannelAttention81.3875.6176.6455.06+2.42 +NegSamples(Final)83.3178.4180.8758.41+4.23 Note:Îvaluesarestep-wiseimprovementsrelativetotheprecedingrow. IntroducingdataaugmentationraisedmAP@0.5to69.46% (+7.31p);addingspatialattention(SAM)increaseditto 74.22%(+4.76p);integratingchannelattention(CAM) broughtitto76.64%(+2.42p);andnegative-sample learningyieldedafinalmAP@0.5of80.87%(+4.23p),an overallimprovementof18.72percentagepointsoverthe baseline.PrecisionandRecallimprovedmonotonicallyfrom 72.28%/67.34%to83.31%/78.41%.Fig.4visualizesthis progressionacrossallfourmetrics. Figure4.Visualizationofdetectionperformanceacrosstheincremental optimizationstages. TheseresultsconfirmthatNMS-freeend-to-end detectioncanbetrainedanddeployedon resource-constrainededgehardwarewithoutsacrificing detectionaccuracy.Losscurvesofboththeauxiliaryand primaryheadsshowedsteadydownwardtrendsandrapid stabilizationacrossthebaselineandthetwooptimization stages(Fig.5).Fielddetectionresultsfromthedeployed JetsonNanoplatformareshowninFig.6. Figure5.Trainingandvalidationlosscomparisonacrossthebaselineandthetwooptimizationstages. Figure6.YOLOv10nweeddetectionresultsinfieldconditions. C.FieldValidation AphysicalprototypeoftheTS-MAMPwasconstructed andevaluatedinfieldtrialsinLu'an,Anhui(Fig.7).The testsconfirmedthestructuralintegrityofthemodulartruss chassisandthereliabilityoftherepurposed(second-life) powertrainunderstandardagriculturalloads.The standardizedinterfacesfacilitatedrapidswitchingbetween functionalmodules.Integratedwiththeoptimized YOLOv10nperceptionsystem,theplatformexhibitedstable mobilityandon-deviceinferencecapability,demonstrating thefeasibilityofrepurposingretiredEVcomponentsfor low-costagriculturalautomation. (a)Physicalprototype(b)Fieldtest Figure7.FieldvalidationoftheTS-MAMPprototype. IV.CONCLUSION ThisworkdemonstratesthatrepurposedretiredEV powertrainscanserveasaviablefoundationforaffordable, intelligentagriculturalmachinery.Byapplying3Rprinciples andscreeningretiredbatterymoduleswithreferenceto GB/T32620.1-2016,weintegratedrepurposed48VBLDC motorsandgraphene-modifiedlead-acidbatteriesintoa functionalplatform,reducingthepowertrain-and-chassis BOMcostbyapproximately60%tobelowUSD450 (perceptionandweedingmodulesexcluded).Onthe perceptionside,theNMS-freeYOLOv10narchitecture enabledon-deviceinferenceontheresource-constrained JetsonNano,yieldinganmAP@0.5of80.87%ontheWanxi Crop-Weeddataset.TheseresultsestablishTS-MAMPasa low-costpathwaythatbridgesindustrialelectronicwaste recyclingwithprecisionfarmingforsmallholderagriculture. Thecurrentapproachoperateswithincertainboundaries, particularlythecapacityheterogeneityofretiredbattery modulesandtheabsenceofmulti-modalsensing.Future workwillprioritizeintegratingtheactivebalancinglogic intoacomprehensiveBMStomanagecapacityvariation, andenhancingenvironmentalrobustnessthrough LiDAR-visionsensorfusionforall-weatherfieldoperation. FUNDING ThisstudywassupportedbytheAnhuiProvincialKey ResearchandDevelopmentProject(2024AH051996),the AnhuiProvincialTeachingReformResearchProject (2023JYXM0695),andtheAnhuiAI+EducationCurriculum Project(2024AIJY327). REFERENCES [1]T.ChamenandJ.McPhee,"Unlockingpotentials:Advantagesof widespancontrolledtrafficfarmingcomparedtostandardCTF," NEXATGmbHCorporateTechnicalNote,2024. [2]O.Mark,"Agritechnica2023:1,100hpNexatgantrysystemto harvest200t/hour,"FarmersWeekly,Nov.2023. [3]V.Bulgakov,S.Pascuzzi,V.Adamchuk,V.Kuvachov,andL. Nozdrovicky,"Theoreticalstudyoftransverseoffsetsofwidespan tractorworkingimplementsandtheirinfluenceondamagetorow crops,"Agriculture,vol.9,no.7,p.144,2019. [4]H.Yangetal.,"Achievingthesustainableagriculturaldevelopment goalsbyadoptingthenewenergyelectricagriculturalmachinery:An analysisofopportunitiesandchallengesofChina,"Energies,vol.18, no.16,p.4211,2025. [5]L.GrimstadandP.J.From,"TheThorvaldIIagriculturalrobotic system,"Robotics,vol.6,no.4,p.24,2017. [6]L.GrimstadandP.J.From,"ThorvaldIIâamodularand re-configurableagriculturalrobot,"IFAC-PapersOnLine,vol.50,no. 1,p.4588â4593,2017. [7]NaĂŻoTechnologies,"TedâWeedingRobot,"2026.[Online]. Available:https://w.naio-technologies.com/en/ted-robot/ [8]FAOInvestmentCentreandInnovationsforPovertyAction(IPA), "Pathwaystoprofit:Experimentalevidenceonagricultural technologyadoption,"FAOInvestmentCentreBrief,Oct.2023. [9]ISFAdvisors,"Beyondthefrontier:Decodingviabilityin smallholderfinance,"2025RuralandAgriculturalFinanceStateof theSectorReport,Nov.2025. [10]T.A.Abetu,P.T.M.Ingenbleek,K.E.Giller,E.Wolde-Meskel,and E.Baars,"Whysmallholders(donot)adoptproductivity-increasing technologies?Developingasmallholderadoption-processmodel throughanin-depthqualitativestudyinEthiopia,"Technologyin Society,vol.85,p.103211,Apr.2026. [11]Xinhua,"ChinaseekstostepupregulationofNEVbattery recycling,"May28,2026. [12]M.N.AkramandW.Abdul-Kader,"Repurposingsecond-lifeEV batteriestoadvancesustainabledevelopment:Acomprehensive review,"Batteries,vol.10,no.12,p.452,2024. [13]E.Antwi,G.K.Ayetor,andF.K.Forson,"Reviewofevaluationand decision-makingframeworkforremanufacturingspentelectric vehicletractionmotors,"Preprints.org,2026. [14]J.G.Erdmann,"Electromobility:Secondlifeforelectricmotors," Fraunhofer-GesellschaftResearchNews,Jan.2024. [15]R.Wang,L.Zhan,Z.Xu,R.Wang,andJ.Wang,"Agreenstrategy forupcyclingutilizationofcorepartsfromend-of-lifevehicles (ELVs):Pollutionsourceanalysis,technologyflowchart,technology upgrade,"ScienceoftheTotalEnvironment,vol.912,p.169609, 2024. [16]M.S.Koromaetal.,"Lifecycleassessmentofbatteryelectric vehicles:Implicationsoffutureelectricitymixanddifferentbattery end-of-lifemanagement,"ScienceoftheTotalEnvironment,vol.831, p.154859,2022. [17]L.C.Casals,B.A.GarcĂa,andC.Canal,"Secondlifebatteries lifespan:Restofusefullifeandenvironmentalanalysis,"Journalof EnvironmentalManagement,vol.232,p.354â363,2019. [18]N.FallahandC.Fitzpatrick,"Exploringthestateofhealthofelectric vehiclebatteriesatendofuse;hierarchicalwasteflowanalysisto determinetherecyclingandreusepotential,"Journalof Remanufacturing,vol.14,no.1,p.155â168,2024. [19]H.Iqbal,S.Sarwar,D.Kirli,J.K.H.Shek,andA.E.Kiprakis,"A surveyofsecond-lifebatteriesbasedontechno-economicperspective andapplications-basedanalysis,"CarbonNeutrality,vol.2,no.1,p. 8,2023. [20]P.Ghisellini,C.Cialani,andS.Ulgiati,"Areviewoncircular economy:Theexpectedtransitiontoabalancedinterplayof environmentalandeconomicsystems,"JournalofCleanerProduction, vol.114,p.11â32,2016. [21]V.M.Scharmeretal.,"Sustainablemanufacturing:Areviewand frameworkderivation,"Sustainability,vol.16,p.119,2024. [22]D.Tiwari,J.Miscandlon,A.Tiwari,andG.W.Jewell,"Areviewof circulareconomyresearchforelectricmotorsandtheroleofIndustry 4.0technologies,"Sustainability,vol.13,no.17,p.9668,2021. [23]Z.Li,A.S.Hamidi,Z.Yan,A.Sattar,S.Hazra,J.Soulard,C.Guest, S.H.Ahmed,andF.Tailor,"Acirculareconomyapproachfor recyclingelectricmotorsintheend-of-lifevehicles:Aliterature review,"Resources,ConservationandRecycling,vol.205,p.107582, 2024. [24]A.Wang,H.Chen,L.Liu,K.Chen,Z.Lin,J.Han,andG.Ding, "YOLOv10:Real-timeend-to-endobjectdetection,"inAdvancesin NeuralInformationProcessingSystems37(NeurIPS2024),2024. [25]GB/T32620.1-2016,Lead-acidBatteriesforElectricRoad VehiclesâPart1:TechnicalConditions,AQSIQ/SAC,Beijing, China,2016. [26]C.Andreasen,K.Scholle,andM.Saberi,"Laserweedingwithsmall autonomousvehicles:Friendsorfoes?"FrontiersinAgronomy,vol. 4,p.841086,2022. [27]P.Zhao,J.Chen,J.Li,J.Ning,Y.Chang,andS.Yang,"Designand testingofanautonomouslaserweedingrobotforstrawberryfields basedonDIN-LW-YOLO,"ComputersandElectronicsin Agriculture,vol.229,p.109808,2025. [28]GB/T7247.1-2024,SafetyofLaserProductsâPart1:Equipment ClassificationandRequirements,StateAdministrationforMarket Regulation(SAMR),Beijing,China,2024. [29]T.P.Swaminathan,C.Silver,andT.Akilan,"Benchmarkingdeep learningmodelsonNVIDIAJetsonNanoforreal-timesystems:An empiricalinvestigation,"arXivpreprintarXiv:2406.17749,2024.