Amid changing traffic lights, simulated rain and fog, and suddenly appearing obstacles,multiple unmanned delivery vehicles navigated the course autonomously at the real-world testing site of the 2025 Guangdong-Hong Kong-Macao Greater Bay Area(GBA) Functional Unmanned Vehicle Competition.

The competition adopted a dual-track model combining specialized testing events with scenario-based excellence challenges. Unmanned delivery vehicles from leading companies including Meituan, JD and Neolix were put to the test across 11 challenges, such as traffic light response, rain and fog recognition, driving on slopes and obstacle avoidance. Hidden tasks including tests involving blocked drivable areas and lane changes across solid lines in construction zones directly targeted safety and efficiency pain points in real-world operations, effectively validating the vehicles' perception, decision-making and control capabilities. A stringent network-based vehicle supervision platform was introduced, requiring fully autonomous driving throughout the competition with real-time status monitoring at three-second intervals.

The competition also explored coordinated scenarios such as "unmanned vehicles plus metro network" and "drones plus unmanned vehicles". In real-world operations, such multi-dimensional logistics networks have already improved efficiency by more than 30% in areas such as Huaqiangbei in Shenzhen's Futian District.

Source:PEARL

全国首个功能型无人车实战大赛深圳鸣枪

信号灯变幻、模拟雨雾弥漫、障碍物突然出现，一辆辆无人物流车自主穿梭其中......这是2025粤港澳大湾区功能型无人车大赛的实景测试现场。

大赛设计了“专项测试赛+优秀场景赛”的双轨模式。美团、京东、新石器等主流企业的无人物流车，直面信号灯响应、雨雾识别、坡道行驶、障碍物规避等十一项挑战。其中，可行驶区域堵塞测试、施工实线变道等隐藏关卡，精准命中实际运营中安全与效率的痛点，有效验证了车辆的感知、决策与控制性能。赛事引入了严格的网联监管平台，要求车辆全程自动驾驶并接受3秒级实时状态监控。

此外，比赛还对“无人车+地铁”“无人机+无人车”等协同场景进行了探索。在福田华强北等区域，类似的立体化物流网络已提升效率超30%。

文、图｜记者 宋王群
翻译｜曾敏
英文审校｜林佳岱