Automation in Construction in brief – 07/03/2026
Automation and Digitization
A new method automates the generation of Building Energy Model (BEM) geometry from Building Information Models (BIM) using minimal inputs-space and opening volumes from IFC files. It bypasses error-prone construction-based approaches, handling complex geometries and scaling efficiently for large buildings. Tested on a university and an office building, the method reduces manual corrections, streamlining energy simulation workflows.
Source: https://doi.org/10.1016/j.autcon.2026.106840
Finnish researchers developed a model-free reinforcement learning (RL) system enabling standard excavators to autonomously capture large rocks. Trained in a high-fidelity simulator with randomized rock properties, the RL agent achieved an 80% success rate on unseen rocks, outperforming traditional model-based approaches. The method avoids explicit modeling of rock or soil dynamics, offering a scalable solution for unstructured environments in mining and construction. Study limited to simulation.
Source: https://doi.org/10.1016/j.autcon.2026.106869
City Detect, an AI startup monitoring urban decay via cameras on municipal vehicles, raised $13m in a Series A led by Prudence Venture Capital. The firm, operational in 17 cities, uses computer vision to track building violations, illegal dumping, and storm damage-processing thousands of cases weekly, far outpacing manual methods. Funds will expand US operations and refine storm-detection tech.
Source: https://techcrunch.com/2026/03/06/city-detect-uses-ai-to-help-cities-stay-safe-and-clean/
Luma, an AI video startup, launched Luma Agents-end-to-end creative tools for text, image, video, and audio, powered by its Unified Intelligence models. The agents automate workflows for ad agencies and brands, generating and refining content via self-critique. Early adopters include Publicis, Adidas, and Mazda. Luma claims the system slashed a 15m $ ad campaign to 20k $ in 40 hours.
Construction Robotics
Qualcomm and German robotics startup Neura Robotics have partnered to develop next-gen robots and physical AI, using Qualcomm’s Dragonwing IQ10 processors. Neura will integrate the chips into its humanoid and industrial robots, leveraging its Neuraverse simulation platform. The deal mirrors trends in robotics, where startups collaborate with tech giants to accelerate deployment. Expect more such partnerships as AI firms target physical AI markets.
Source: https://techcrunch.com/2026/03/09/qualcomms-partnership-with-neura-robotics-is-just-the-beginning/
A vision-driven robotic system is transforming wood-framed construction, achieving 96.9% grasping accuracy and millimetre-level precision without CAD models. By using real-time perception and adaptive control, the technology overcomes wood’s natural variability, slashing errors by 1.8-6.1× compared to traditional methods. This breakthrough could automate small-batch construction, boosting efficiency in an industry still reliant on manual labour.
Source: https://doi.org/10.1016/j.autcon.2026.106858
Zoox expands robotaxi mapping to Dallas and Phoenix, deploying Toyota SUVs to gather data ahead of autonomous testing. The Amazon-owned firm will later switch to its custom driverless vehicles, pending federal approval. With these additions, Zoox will operate in 10 U.S. cities, leveraging Arizona’s AV-friendly regulations. The move creates jobs and bolsters its fleet of 300,000+ passenger rides.
Source: https://techcrunch.com/2026/03/09/zoox-starts-mapping-dallas-and-phoenix-for-its-robotaxis/
BYD’s flash-charging EV battery cuts recharge time to five minutes-with a catch.The Chinese automaker’s new Blade Battery 2.0, debuting in the Yangwang U7 sedan, charges from 10% to 70% in five minutes but requires proprietary 1.5MW chargers. While LFP chemistry reduces costs, range lags behind rivals like Lucid. BYD aims to expand its 4,200-strong charging network to 20,000 by year-end, seeking an edge in China’s fierce EV price war.
Space Construction
NASA’s Artemis II, a 10-day crewed lunar mission, will test Orion’s deep-space systems ahead of future Moon and Mars ventures. The public can monitor its progress via AROW, a real-time tracker showing Orion’s location, speed, and trajectory-on NASA’s site or its app, which includes augmented reality. Data, including state vectors and ephemeris files, will be open for third-party visualizations. A stepping stone to sustained lunar exploration. The mission will start earliest on march 6.
Source: https://www.nasa.gov/missions/artemis/artemis-2/track-nasas-artemis-ii-mission-in-real-time/
The European Space Agency is accepting applications for its Disruptive Innovation and Commercialisation in Earth Observation course (6-9 July, Belgium). Following a successful 2025 pilot, the programme-run with ESA Φ-lab-targets engineers, scientists, and entrepreneurs, blending theory with hands-on training to bridge deep tech and business. A chance to turn space data into market-ready solutions.