机器人“监工”上线

“Hey, look busy. Our robot boss is trundling over here!” That’s a sound that you may very well hear on the construction sites of the future, thanks to a lidar-equipped robot and matching drone developed by the company Doxel. Emerging from stealth this week with a $4.5 million funding round, Doxel hopes to revolutionize the construction industry using robots to scan construction sites to make sure that work is progressing correctly.

“嘿，快忙起来，我们的机器人老板正在附近视察呢！”未来的建筑工地上很可能会经常听到这种声音，这要得益于Doxel公司开发的一款内置激光雷达的机器人和配套无人机。曾出产隐形飞机的Doxel本周获得4500万美元投资，希望在建筑行业掀起革命，用机器人扫描工地情况，确保工程正确进行。

Far from a simple robotic camera, however, Doxel’s smart robots harness cutting-edge deep-learning technology to make informed judgements about how well work is going. If things fall behind schedule, a construction manager will receive a smartphone notification reading, “Mechanical team productivity has dropped below threshold” or similar. In other words, simply trying to look busy on site won’t be enough to fool these smart management bots.

Doxel公司的智能机器人远不止机器摄像头这样简单，它借助前沿深度学习技术在了解现场情况后可以很好地判断工程进度。如果工程进度落后，建筑经理会在智能手机上收到类似“机械团队产率低于门槛值”之类的提示。换言之，在工地上装样子的人没办法糊弄智能管理机器人。

“We make use of drones and rovers as data-capture mechanisms; drones for outdoors and rovers for indoors,” Saurabh Ladha, CEO and co-founder of Doxel, told Digital Trends. “The vast majority of construction involves indoors and that is where we deploy lidar in addition to an HD camera. The lidar captures depth data. Once the data is captured, our deep learning kicks in to automatically inspect quality and measure installed quantities. The diversity of objects found in construction and the degree of noise, clutter, and occlusion you find in such environments had made the automation of this process nearly impossible. Doxel’s A.I. capabilities means that job sites are automatically digitized, that project quality undergoes ongoing inspection, and that progress is measured in real time. Our dashboard then offers automatic real-time budget, schedule, and quality reports. We help to keep molehills from becoming mountains.”

“我们使用无人机和探测车作为数据捕捉机制；无人机用于室外，探测车用于室内”Doexl公司首席执行官和联合创始人绍拉巴·拉德哈告诉Digital Trends网站“绝大多数建筑工程都在室内进行，除高清摄像头外，我们还会安装许多激光雷达。一旦捕捉到数据，深度学习自动监测工程质量，估计工程进度。工地现场材料物品繁多，封闭又杂乱的环境十分嘈杂，实现自动化几无可能。Doxel公司的人工智能技术意味着工地现场自动实现数字化，工程质量可进行持续监测，工程进度也可进行实时估计。随后屏幕会自动提供实时预算，进度表与质量报告。这有助于企业防微杜渐。”

Ladha observes that many construction projects — as many as 98 percent of large ones — can go significantly over time and budget. Being able to spot these trends early means that changes can be made in order to ensure things progress as efficiently as possible.

拉德哈注意到许多建筑工程，即多达98%的大型建筑工程都存在严重的超支超时问题。及早发现超支超时的趋势意味着可以做出改变确保工程尽可能高效地进行。

“We are currently working with large construction companies such as DPR Construction, as well as a large healthcare company building hospitals and medical office buildings,” Ladha said. “We sell Doxel as a service: everything from capturing data to delivering insights, and the service is sold on a monthly fee basis. Having just come out of stealth, our plans are to grow the business. The construction industry has struggled with productivity and we believe we can help to turn that around.”

“目前我们正和DPR建筑公司这样的大公司展开合作，同时我们的客户也包括一家正在建设医院和医疗办公楼的大型保健公司”拉德哈表示“Doxel产品的卖点在于服务：从捕捉数据到提供分析应有尽有，按月收费。从隐形飞机领域转型后，公司计划发展这项业务。建筑行业一直受困于产率问题，我们认为自己足以帮助建筑行业解决这一问题。”