[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-method-builds-route-planning-into-self-driving-ai-models":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},6725,"new-method-builds-route-planning-into-self-driving-ai-models","New Method Builds Route Planning Into Self-Driving AI Models","A new technique called DiffAdapterVLA folds trajectory planning directly into a driving AI's neural layers, cutting latency without a separate planner.","A new architecture teaches self-driving AI to plan its route while it thinks, not after.\n\nResearchers built DiffAdapterVLA, a method that injects trajectory generation into the late layers of vision-language driving models, rather than bolting a separate planning module on after the model finishes reasoning. It works through lightweight \"DiffAdapters\" attached layer by layer, letting a trajectory refine itself repeatedly as data moves through the network, steered by driving context through what the researchers call asymmetric joint attention. Because it reuses the model's existing layers instead of training a new planner from scratch, the approach only needs to fine-tune a small number of parameters. On the NAVSIM benchmark, the team reports high-quality closed-loop planning with low end-to-end latency.\n\nMost driving vision-language models treat scene understanding and route planning as separate jobs, with a planner tacked onto whatever the language model outputs at the end. Merging the two inside the same forward pass is a real efficiency argument: fewer components to train, run, and keep in sync, which matters when a car has milliseconds to decide what to do next.\n\nStill, NAVSIM is a simulation benchmark, not a road test, so how this holds up in messy, real-world driving is an open question.","[\"autonomous driving\",\"vision-language models\",\"ai research\"]","2026-09-17T04:00:00.000Z","2026-09-18T07:51:21.473Z","2026-09-18T07:51:33.460Z","published",null,[],"ai",[26,27,28],"autonomous driving","vision-language models","ai research",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.15322",0,{"sections":35},[36,40,44,49,54,58,62,67,72,76,81,86,91,96],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",3852,"2026-09-17T08:27:09.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",648,{"name":45,"slug":46,"count":47,"latest_published_at":48},"Policy","policy",338,"2026-09-11T04:00:00.000Z",{"name":50,"slug":51,"count":52,"latest_published_at":53},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":55,"slug":56,"count":57,"latest_published_at":18},"Hardware","hardware",154,{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",114,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",99,"2026-09-09T17:27:33.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Software","software",75,"2026-09-10T20:41:21.000Z",{"name":73,"slug":74,"count":75,"latest_published_at":18},"Dev Tools","dev-tools",73,{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",55,"2026-09-09T23:14:29.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",43,"2026-09-10T12:18:06.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",41,"2026-09-08T01:57:23.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]