Rayfast is a leading stocking distributor of quality interconnect solutions as well as high-performance harness components and electro-mechanical products. Key markets include aerospace, defence, energy, industrial, rail and motorsport, which are supported by an experienced team, with technical specialists by sector and product.

Weald Electronics is known for its comprehensive selection of circular bayonet and screw coupling connectors such as the LMH, LMJ, LMV and LMG series, produced in a choice of materials and finishes and leads the way in high performance interconnection solutions based on popular industry standards including industry standard MIL-DTL-5015 and MIL-DTL-26482 specifications. The versatility of Weald’s circular connectors is facilitated by a wide choice of shell sizes, styles, plug and socket contact types and contact arrangements. 

As well as circular connectors, Weald Electronics manufactures two-part PCB connectors and subminiature plastic-bodied circular connectors and a comprehensive range of connector accessories, including high-performance and ultra-lightweight protective covers, nut plates and gaskets specifically designed for Motorsport and UAV applications.

Rayfast will have full access to the complete range of Weald products, which includes:

• Electrical connectors
• Backshells and adaptors
• Protective caps and dust covers
• Connectors accessories.

Nick Wheeler, sales director for Weald Electronics says: “With their established and highly regarded presence in our industry, Rayfast will provide an excellent route to market for Weald Electronics, especially for customers who are looking for supply as part of a wider Harness Components portfolio of products, such as wire and cable, markers, tubing and heatshrink moulded shapes. The Weald product compliments their existing connector ranges and they have a superb team that is dedicated to sales, technical support and customer service”

Neil Ruddick, UK sales director for Rayfast, commented: “We are very excited to agree what I am sure will be a long-term partnership with Weald Electronics. Their excellent product range, capabilities and reputation for quality & customer service fit perfectly with our core strengths in electrical interconnect, providing even greater depth and breadth of solutions for our customers.”

www.wealdelectronics.com

Combining 15.0A power and high-speed Ethernet communication within a single, ultra-compact interface, MiniMix streamlines installation in space-constrained joints, actuators and robotic mechanisms, reducing connector count, cabling complexity and assembly effort.

“Humanoid robot manufacturers are rapidly moving from prototype builds to commercial mass production, but traditional joint wiring methods create major manufacturing bottlenecks," said Brian Hauge, president and SVP, consumer and commercial solutions, Molex. “MiniMix replaces tedious manual assembly with an integrated solution for power and signal transmission, freeing robotics designers to build sleeker, lighter and more responsive autonomous systems at scale.”

Integrating power and signal connectivity into increasingly compact joints, actuators and articulated mechanisms without sacrificing performance is a major hurdle for the robotics industry. Design engineers face severe space and assembly constraints in multi-axis articulated joints – such as wrists, elbows, knees, ankles and neck assemblies – where actuator motion systems must route high power and high-speed control data through extremely tight channels. 

Traditional approaches using separate power and signal connectors are typically too bulky and heavy to fit in these confined spaces, forcing manufacturers to thread bare wires manually and hand-solder terminals during final assembly, which introduces potential failure points and slows production output.

The Molex MiniMix Hybrid Power and Signal Connectors resolve these manufacturing challenges by consolidating up to 15.0A power contacts and 1Gbps 1000BASE-T1 Ethernet communication into a single integrated interface. Featuring an ultra-compact 5.65mm routing profile, MiniMix is the industry’s smallest hybrid connector, requiring up to 50% less routing area than alternative options. This enables pre-terminated cable assemblies to pass smoothly through narrow internal actuator channels, articulated limbs, robotic joints and end-effectors. 

Offered in vertical and right-angle mating orientations, the Molex MiniMix Hybrid Power and Signal Connectors can accommodate diverse layout requirements while optimising space utilisation and reducing cable strain. Furthermore, its robust, vibration-resistant mechanical design withstands continuous motion and constant flexing, ensuring reliable electrical performance and long-term durability across repeated movement cycles.

The Molex MiniMix Hybrid Power and Signal Connectors offer wire-to-board connector and cable assembly solutions that are ideally suited for evolving robotic applications where space, weight and reliability are crucial. According to a 2026 supply chain analysis by McKinsey & Company, joint actuator systems represent 40% to 60% of a humanoid robot’s total bill of materials (BOM), intensifying requirements to optimize sub-assembly space and eliminate manual cabling labor. As robots scale to include approximately 50 articulated joint actuators, integrated connectivity solutions like the Molex MiniMix Hybrid Power and Signal Connectors will play a pivotal role in facilitating plug-and-play assembly, reducing harness weight and minimizing potential failure points across next-generation humanoid robotic architectures.

Samples and functional evaluation assemblies of the Molex MiniMix Hybrid Power and Signal Connector family are available now for customer design-in and joint evaluation, with full commercial production scaling slated for late 2026.

www.molex.com

The report reveals that as manufacturers increase their AI investments, the focus is shifting from standalone automation projects to larger physical AI ecosystems. It highlights that companies see physical AI as a people-first transformation. Instead of replacing the workforce, manufacturers are using intelligent systems to help employees work more safely, efficiently, and productively while supporting workforce redeployment where needed in a Human + AI Operating model.

The report also highlights how enterprises are preparing to scale physical AI across factories, warehouses, logistics networks, maintenance operations, and quality management environments. It provides insights into the strategies, challenges, and opportunities involved in adopting physical AI at scale. The study builds on TCS’ partnership with Google Cloud, following the March 2026 launch of the TCS Physical AI Gemini Experience Center in Troy, Michigan. The center helps manufacturers explore, test, and scale physical AI use cases for safety, quality, and operational efficiency with intelligence at the core.

The study surveyed CXOs and vice presidents from 300 manufacturing companies across North America and Europe between March and April 2026. Respondents represented sectors such as automotive, electronics and high-tech manufacturing, industrial equipment and machinery, process industries, and aerospace and defense.

Key findings from the study include:

• Manufacturers are committing to long-term investment: No surveyed organisation plans to reduce Physical AI investment, while 26% plan to increase spending. Manufacturers are preparing for longer value-realization timelines, underscoring sustained transformation over short-term pilots.
• High impact expected in warehouse operations: 77% anticipate significant or transformational impact, followed by assembly and manufacturing operations at 75% and logistics and material movement at 72%.
• Human-centric transformation is taking shape: 42% of manufacturers expect significant workforce augmentation through Physical AI by improving safety and supporting workers in complex, hazardous, or repetitive industrial environments.
• Early-stage adoption, clear roadmap to scale: 68% of manufacturers remain in non-deployment or experimental stages, with legacy system integration, data infrastructure, and workforce skills identified as the primary paths to enterprise-scale deployment.
• Governance is the next frontier: 44% report unclear or no formal accountability structure for Physical AI failures, and 40% are unprepared for emerging regulatory requirements, signaling where manufacturers should focus as deployment accelerates.

Anupam Singhal, president, manufacturing, TCS, said: “Physical AI is taking intelligence beyond the screen and onto the shop floor, where machines sense, adapt and act in real time. The manufacturers that scale it successfully will define the next era of manufacturing. TCS’ ‘infrastructure to intelligence’ approach positions them to lead that transformation. With all the manufacturers in our study planning to either maintain or increase the investment, the direction is clear: towards more resilient, adaptive, and future-ready manufacturing enterprises.”

Kevin Ichhpurani, president, global partner ecosystem, Google Cloud, said: “Physical AI is moving manufacturing from digital insight to autonomous real-world action. Through our partnership with TCS, we are bringing Gemini’s multimodal reasoning to the factory floor, enabling robots and systems to operate safely and intelligently in complex industrial environments.”

TCS is helping global manufacturers move from pilots to enterprise-scale physical AI adoption by bridging the gap between digital ambition and physical deployment. This is enabled through the TCS Physical AI Gemini Experience Center, in Troy, Michigan. Similarly, the TCS Physical AI Blueprint offers an end-to-end framework that integrates AI-powered quadruped and humanoid robotics with advanced sensing, edge intelligence, and secure cloud orchestration to deliver real-time operational insight and autonomous decision support.

These capabilities, combined with TCS’ deep manufacturing domain expertise, help enterprises build scalable, governed, and future-ready Physical AI ecosystems.

www.tcs.com

Delivered through the UK's Modern Industrial Strategy, the funding will support over 1,800 jobs - and thousands more in the supply chain - across the UK, helping to put more money in people's pockets and ensure Britain is better off.

Auto firms and key R&D partners have been awarded nearly £50 million in government funding that will help businesses build and produce the zero-emission vehicle technologies of the future at scale - helping to drive economic growth while simultaneously making EVs cheaper for consumers.

This latest funding is delivered through the DRIVE35 programme - the biggest government investment into the UK’s car industry of the post-war era, with £4 billion to 2035 speeding up the electrification of the automotive industry. This builds on the Government’s mission to reindustrialise every part of the UK, back British innovation and ensure companies are able to succeed and scale here at home.

Industry Minister Blair McDougall said: "Britain invented the modern motor industry and we're determined to ensure the next generation of vehicles are designed and built here too. This investment will secure skilled jobs, strengthen our manufacturing heartlands and help drive the reindustrialisation of Britain. Through our Modern Industrial Strategy we're backing working people and British businesses to deliver growth in every part of the country."

The Government is doubling down on its commitment to delivering growth in every region. With the North East and West Midlands serving as the UK's powerhouse regions for automotive manufacturing, this funding will provide vital support to flagship companies such as Turntide Technologies and Bentley, helping to ensure the next generation of vehicles are designed, developed and built here in the UK.
 
This follows £100 million worth of DRIVE35 grant funding already being delivered to supply chain businesses in both of these regions, preparing them for the electric vehicle transition. With the connected and automated mobility (CAM) industry expected to create tens of thousands of high-quality jobs and contribute around £66 billion to the economy by 2040, according to SMMT. The Government is injecting £17 million into a further nine successful CAM Pathfinder projects, paving the way for innovators to develop, test and scale cutting-edge transport technologies - reinforcing the UK's position as a global leader in this growing industry. Altogether this industry-matched funding totals nearly £130 million, showcasing the best of British innovation on the global stage and bolstering the UK’s advanced manufacturing sector.

Ian Constance, Chief Executive at the Advanced Propulsion Centre UK, said:  “The projects announced today demonstrate the depth of innovation and engineering excellence that exists across the UK automotive sector. Through DRIVE35, we are supporting businesses to move promising technologies to commercial deployment and manufacturing at scale.

“This investment is about far more than individual projects. It is about strengthening the UK's capability to design, develop and build the technologies that will define the vehicles of the future. By bringing together industry, government and academia, DRIVE35 is helping create the conditions for long-term growth, increasing investor confidence and reinforcing the UK's position as one of the world's leading destinations for automotive innovation."

www.gov.uk

By bringing drives, motors, gearboxes, control equipment and PLCs together under one manufacturer, Technidrive can deliver complete, end-to-end systems that reduce integration complexity while improving visibility, connectivity and support.

The move is particularly valuable for applications where equipment performance must be monitored continuously, but reliable internet access cannot be guaranteed. This includes remote or mobile assets such as marine systems, where data must be captured locally and synchronised later so operators can retain a complete record of asset condition.

As a WEG Premier Partner, Technidrive already supplies the Brazilian manufacturer’s motors, variable speed drives (VSDs), gearboxes and control equipment. By adding WEG PLCs, the company can deliver fully integrated automation systems on a single control architecture, improving communication, reducing complexity and streamlining commissioning and support.

Marine data acquisition

This approach has already been applied in a marine data acquisition project, where the customer needed a reliable way to monitor onboard assets despite inconsistent connectivity at sea. Technidrive deployed the PLC500 to track temperature, humidity and pressure across large vessels, storing data locally and syncing it when a connection is available. This allows the customer to retain a full record of asset conditions, so engineers can later identify environmental issues, check for pressure drops, spot early signs of deterioration and plan maintenance before faults cause failure or downtime.

“The ability to bring everything together under one manufacturer is a major advantage,” explained David Strain, technical director at Technidrive. “Knowing the PLC has been tested alongside the drives, motors and control gear gives us confidence in system performance and reliability, while also helping to ensure smooth integration, faster installation and simpler support for our customers.”

A key benefit of the PLC500 in this application is its remote connectivity. For vessels operating across different regions, internet access can be inconsistent or restricted. By storing information locally and uploading it once a connection becomes available, the PLC500 enables operators to maintain a reliable record of system performance over time.

Multiple languages

Technidrive also highlighted the PLC500’s programming environment and human-machine interface (HMI) as important factors in its adoption. Built on the widely recognised CODESYS platform, the PLC supports multiple programming languages, including structured text, ladder logic, function block, sequential function chart and continuous function chart. This gives engineers the flexibility to work with familiar tools and adapt the system to different applications.

“The software and ease of development were major drivers for us,” added Strain. “Our engineers can program in multiple languages, and the system is very intuitive to use. Combined with strong connectivity and remote access, it gives us a flexible platform for a wide range of applications.”

With support for multiple communication protocols and networking options, the PLC500 can integrate with a wide array of devices and systems, making it suitable for industries ranging from marine and materials handling to quarrying, recycling, food processing and general manufacturing.

As part of WEG’s wider PLC portfolio, which includes models such as the PLC410 for medium-sized applications, the compact PLC20X for smaller systems and the PLC300 designed for operator interface and machine-level control, the PLC500 sits within a scalable ecosystem tailored to applications of varying size and complexity.

“PLCs are inherently flexible, so the opportunities are wide open,” said Strain. “The experience has been very positive, and WEG’s PLCs give us a strong platform for future automation projects.”

technidrive.co.uk

Last week NVIDIA announced Halos for Robotics, which it describes as the industry’s first full-stack safety system for physical AI. The system brings the company’s autonomous vehicle safety work into the world of robots, spanning computers, sensors, software and a new inspection lab to help partners prepare for third-party certification. Agility Robotics is the first to build NVIDIA Halos for Robotics into its humanoid robot, Digit.

This is good news, and I want to say so plainly. Safety has long been the least glamorous corner of robotics. It appeals to engineers and standards committees, not to the people who make the videos of robots dancing and running half marathons. So, when one of the most influential companies in computing puts its weight behind robot safety, and when a leading humanoid maker submits its system for independent scrutiny, the whole field benefits.

Investment in safety is investment in trust, and consumer confidence is critical if businesses are to continue to benefit from automation. It’s worth looking closely at what this wave of work addresses, and at what it does not.

A perception problem, well funded

Most of the safety effort around humanoids, including much of the new tooling, concentrates on perception. The hard questions are framed as “can the robot detect a person, predict where they are going, and avoid a collision?” NVIDIA’s approach includes an ‘outside-in’ method that uses external cameras and AI agents to watch the workspace and adjust how the robot behaves. That is difficult but crucial engineering.

However, perception-based collision avoidance rests on an assumption that the robot remains in control of itself. The discipline of functional safety is built on the opposite instinct. A risk assessment spends less time on how a machine behaves when everything works, and more on how it behaves when something breaks. The central question is what the robot does the moment a part fails – in the case of humanoids the answer is uncomfortable.

What happens when a part fails

A legged humanoid is dynamically stable. It stays upright only by continuously sensing its own balance and correcting, many times a second. Take that control away and it falls rather than pausing in place.

Picture a humanoid mid-stride when an actuator in one leg fails. The machine is already committed to a step it can no longer complete, and it topples, potentially onto whoever is nearby. Given the mass of these robots and their high center of gravity, the potential consequences are serious. A flawless perception system is no help here – the robot might know where people are but will be unable to avoid them. By the time the fault occurs, the robot has lost the very control that perception was meant to inform.

My colleague and leading robotics safety expert, Roberta Nelson Shea, makes a related point about power: a battery fault can disable every actuator and sensor at once, and a humanoid needs a reserve of power simply to lower itself to a stable position. Stability depends on nothing going wrong which makes fault tolerance the deeper safety challenge for human-shaped robots.  

Where this meets the rule book

Industrial robots are typically designed to be single-fault tolerant to meet the relevant safety standards. In plain terms, no single component failure should be able to create a dangerous situation. ISO 10218-1, the standard for industrial robot arms, is built on that principle.

Apply the same requirement to a heavy, dynamically stable machine and you are left with two options. The first is to engineer actuators, sensors and battery management that are either fully redundant or so reliable they almost never fail – how to do that at an acceptable cost currently eludes me.

The second option is to accept that working alongside a humanoid carries more risk than working alongside a conventional industrial robot, because a single failure can lead to a hazard.

ISO 10218-1 does not address the risks created by a humanoid’s own mobility. The first standard written specifically for dynamically stable mobile robots, ISO/CD 25785-1, is still at committee-draft stage, working through its first round of comments. As a first attempt at hard new ground, it has a long way to mature.

None of this means humanoids are impossible. It simply means the safety case for them is unfinished, and the honest version of that case is more demanding than the current conversation suggests.

The advantage of being purpose-built

A mobile collaborative robot, a robotic arm mounted on a wheeled autonomous base, does not face these problems because it is statically stable – a heavy base, a low center of gravity, and no need to balance. When something goes wrong, the safe response is simple, and crucially it is simple to make redundant. Cut the power and apply the brakes. A machine standing still is rarely a danger to anyone.

That difference is the result of choosing a form factor for the task rather than for its resemblance to us. Autonomous mobile robots, collaborative arms, and the standards-driven safety functions behind them were designed so that the failure modes are predictable and the safe states are easy to reach. The robot’s safety does not hinge on a perception system continuing to work. It is built into the architecture before any software runs.

There is another reason that purpose-built automation will keep winning on the factory floor. The complexity that makes a humanoid captivating in a demonstration is the same complexity that multiplies its failure modes. In manufacturing, reliability is non-negotiable, and falling short means downtime, rework and risk.  

Humanoids are not the only way to harness the incredible potential of physical AI on the factory floor. A rapidly growing number of AI enabled applications are running on collaborative industrial robots and autonomous mobile robots handling machine tending, palletising, inspection, and material movement with the safety and reliability the factory floor demands.

The question to ask first

I welcome NVIDIA’s move, and I hope it marks the start of a more serious, better-resourced era for robot safety. The more rigor the industry brings, the better for every company building machines that share space with people, mine included.

But anyone evaluating a humanoid for real work should start with the important unglamorous question. Before asking whether the robot can recognise a worker, ask what it does when an actuator, a sensor or a battery fails mid-task. Ask for the stability and reliability data. Ask which standards the system is certified against, and which parts of its operation those standards actually cover.

We had this debate, rightly, about self-driving cars, an open and public discussion about how much risk is acceptable and who decides. Humanoids deserve the same scrutiny. Until the field can answer the failure question as confidently as it answers the perception question, the proven path remains the one the industry has been building for years: robots whose shape, and whose safety, are designed for the job.

David Brandt is VP of R&D and CTO at Teradyne Robotics 

www.teradyne.com/robotics/

Figures derived from the company’s latest market study, “The Global Industrial Wireless Solutions Market”, show annual shipments are expected to grow at a compound annual growth rate (CAGR) of 8.1% to reach 8.5 million by 2030.

Although wired networking solutions are still predominantly used for industrial communications between sensors, controllers and systems, wireless solutions are widely used as wire replacements in hard-to-reach or hazardous areas, on moving machine parts and on portable equipment.

“The standardised wireless technologies Wi-Fi, IEEE 802.15.4 and Bluetooth have advanced to become the leading wireless technologies for industrial applications”, said Veronika Barta, IoT analyst at Berg Insight.

Other wireless technologies include for example cellular, proprietary radio solutions in the unlicensed ISM 900 MHz and 2.4 GHz bands, non-3GPP LPWA, satellite, IO-Link Wireless and non-802.15.4 WPAN technologies.

Today, a growing number of wireless field devices are provided by many major industrial automation vendors including ABB, Emerson, Honeywell Technologies, OMRON, Rockwell Automation, Schneider Electric, Siemens and Yokogawa. Instrumentation is particularly important in process industries where continuous monitoring and precise control of process variables are essential. Field instruments are also used in factory automation applications such as machine and condition monitoring. Suppliers of industrial wireless sensing solutions for process and factory automation include Balluff, Banner Engineering, Endress+Hauser, ifm electronic, KROHNE, Pepperl+Fuchs, SICK, Turck and VEGA.

Major providers of wired industrial network equipment also offer wireless solutions to enable customers to monitor and control devices wirelessly in parts of the plant that are normally not connected due to accessibility or wiring costs. These include Siemens, Cisco, Belden, Moxa and Phoenix Contact, which all offer industrial wireless devices such as routers, gateways and access points along with their wired solutions. Cellular solutions are typically used for data acquisition and backhaul communications in distributed automation applications.

The largest providers of cellular IoT gateways and routers in the industrial space include Teltonika Networks, Cisco, Moxa, Siemens, HMS Networks, Phoenix Contact, Advantech, Robustel, Digi International and InHand Networks. Besides the leading industrial WLAN solution suppliers Cisco, Belden, Siemens and Moxa, additional providers of WLAN devices include Phoenix Contact, Westermo (Ependion), Advantech, HMS Networks and ACKSYS (iXnov Group). 

“Cybersecurity has become one of the highest strategic priorities for suppliers of industrial networking and control solutions as environments become increasingly connected," said Barta.

The convergence of OT and IT combined with the growing adoption of industrial IoT, remote access and edge computing has significantly expanded the attack surface of industrial systems. Device vendors are progressively embedding cybersecurity capabilities directly into their products. End users increasingly expect industrial solutions to provide continuous monitoring and lifecycle management.

“Cybersecurity is no longer viewed as a standalone technology layer but rather as a foundational design principle across the industrial automation system architecture," concluded Barta.

The ability to deliver robust cybersecurity capabilities has thus become an important differentiator for solution providers, particularly in sectors such as manufacturing, energy, utilities and critical infrastructure.

www.berginsight.com

The company is developing intelligent robots that anyone can train and use. Instead of requiring specialist programmers whenever a task or production process changes, workers can show the robot what to do, watch it attempt the task and correct it on the spot.

“A useful robot should be able to learn from the person doing the work,” says Jonathan Scholz, co-founder and CEO of Reimagine Robotics. “They should be able to show it a task, put it right when it makes a mistake and move on to the next problem. That is what it means for a robot to learn on the job. It’s a process we call ‘monkey-see, monkey-do’.

“A robot should arrive with the attitude of a new colleague: ‘How can I help? What do you want me to do?’ The people who understand the process should be able to answer those questions by showing the robot directly.”

Scholz founded Google DeepMind’s Applied Robotics team in London and led it for seven years. He later co-founded Reimagine Robotics in April 2025 alongside former colleagues Oleg Sushkov, Akhil Raju and Misha Denil, with headquarters in London and Sydney. That first phase was funded by pre-seed financing from venture capital firms Fly Ventures and firstminute capital and a number of angel investors.

The next phase

“We’re excited to come out of stealth,” Scholz says. “This last year was about building a core product and a team, and working with customers to test the platform. After working with several partners, we’re absolutely convinced it’s not only viable, but that there is a need for this technology, and that it has massive potential across an array of industrial and manufacturing settings.”

“This next stage is about a new round of fundraising, expanding our team for more deployment muscle, putting robots into more workplaces, and showing that each deployment can make the next one faster, more reliable, and more efficient.”

Robots already on the factory floor

The company is already deploying its robots in advanced manufacturing and electronics disassembly facilities. At a made-to-order plastics business, Reimagine Robotics trained its robots to tend 3D printers overnight by removing print beds, operating latches and pressing controls. The customer’s own team used the platform to automate additional stages, including washing, curing and drying.

In a separate deployment involving the recovery of valuable critical materials from used hard drives, Reimagine Robotics worked with process engineers to develop a three-robot disassembly cell. The resulting workflow combines robots and people working together to evolve and optimise the workflow in real-time.

During the project, Reimagine Robotics reduced the time required to prototype and test a new robot behaviour from approximately one day to around ten minutes. That allowed the robots to become part of the process-design conversation: teams could propose a new use for a robot and test the idea almost immediately.

Humans at the heart of robotics

Throughout all Reimagine Robotics deployments, people remain central to the model.

“For us, this is not about taking people out of the process,” Scholz says. “A robot that learns on the job depends on people. The worker identifies the bottleneck, shows the robot how to help and corrects it until it is useful.

“The robot turns that person’s knowledge into leverage. Instead of someone having to repeat a tedious physical task thousands of times, they can teach the robot, and apply that ability wherever it is needed. I think of it more as a tool to amplify human labour.”

The cash boost will fast-track the development of automated technology that can do everything from planting seeds to picking fruit, easing the pressure on farms that struggle to find enough seasonal workers at harvest time. 

The latest cash injection from the government’s Farming Innovation Programme forms part of a recent £53 million boost and takes total funding available this year to £123 million for cutting-edge farming research and technology. 

Companies such as Fieldwork Robotics are already developing robots that are transforming English farming. The firm has built a raspberry-picking robot and is now using government funding to make it tougher and more reliable so it can be rolled out commercially. 

Chris Danks, head of agrifood at Innovate UK, said: "Robotics and automation are becoming increasingly important tools to help farmers improve productivity and build more resilient businesses. Working in partnership with Defra, we’re supporting businesses, researchers and farmers to work together to develop technologies that address real-world challenges across agriculture. This competition will help accelerate innovations from concept to on-farm application, supporting solutions that can improve productivity, sustainability and resilience across the sector."

This round is open to livestock applications too, building on work such as Roboscientific’s DETECT project to develop technology that can sniff out illness in dairy cows before it takes hold. Using breath analysis, the system aims to catch bovine respiratory disease early, an illness that costs farmers dearly in lost animals, vet bills and extra work. 

It also casts the net wider than the previous automation round which ran in 2023, welcoming applications from ornamental plant propagation, growing and monitoring, where the technology could equally be put to work in food-producing horticulture, as well as forestry projects covering tree nursery and planting operations and woodland monitoring and management systems. 

Delivered in partnership with Innovate UK, the government’s Farming Innovation Programme backs bright ideas that solve real problems on English farms. The programme is part of a government pledge to invest at least £200 million in farming innovation by 2030. 

Innovators can find out more and apply through the Innovation Funding Service website.

Image: MIT researchers

This technique enables the artificial intelligence model that plans a robot’s motion to forecast its future position. The model uses this prediction to seamlessly transition current movements into the next actions.

Many existing methods cause a robot to stop and think about what it needs to do next, leading to slow and jerky motions. By basing its calculations on the future state of the robot, rather than its current position, the MIT method helps robots operate much faster. Importantly, the technique does not add any computational overhead to the planning process and can be applied to varied robotic hardware.

This new method doubled the speed of robots performing activities like pick-and-place tasks, while significantly reducing lag time between motions. It also boosted the performance of robotic arms in highly dynamic activities, such as playing table tennis and Whack-a-Mole.

The system could be especially useful for robots that perform fast and agile manoeuvers in challenging real-world environments, like emergency response or search-and-rescue. It could also allow robots to react more quickly when recovering from mistakes.

“This work sets up a good foundation for efficient, fast, accelerated, and low-cost robotics applications. We look forward to expanding our work into the latest world action models, so it has even stronger capabilities as we keep pushing to make physical AI faster,” says Song Han, an associate professor in the MIT Department of Electrical Engineering and Computer Science (EECS), member of the Research Laboratory of Electronics, and lead author of a paper on this method.

Han is joined on the paper by co-lead authors Jiaming Tang, an MIT EECS graduate student, and Yufei Sun, a student at Tsinghua University; as well as others at Nvidia, the University of California at Berkeley, the University of California at San Diego, and Caltech. The research will be presented at the Intelligent Robots and Systems Conference. 

Forecasting the future 

In state-of-the-art robotics applications, generative AI systems called vision-language-action (VLA) models act as the brain of a robot, planning its next moves and executing those actions. 

A VLA model takes environmental observations from the robot’s camera and instructions about its task, outputs the next few motions as one chunk of actions, then executes those actions on the robotic hardware.

But VLA inference — the real-time procedure during which the model processes visual inputs, reasons about the task, and outputs actions — is computationally demanding, so the robot can experience substantial pauses while planning its next actions. These pauses disrupt the fluidity of its motions and make it slower to react to changes in the environment. 

“Our motivation was to overlap the thinking process with the execution process to make the reaction speed faster,” Tang says.

The MIT researchers developed a new system called VLASH that enables a VLA to predict the future state of the robot and its environment. It uses this information to plan the next set of motions while the robot is completing the current action chunk.

This solves a major hurdle faced by many other methods, which use the current state of the robot to predict its next moves.

“Since the environment will change after the robot moves, if we plan based on stale observations of the current environment, there will be a misalignment that causes very unstable control,” Tang explains.

VLASH avoids this misalignment due to a key insight by the researchers. Although the model doesn’t know exactly what the environment will look like in the future, it does know the robot’s current position and how it will move to perform the actions it is about to take. 

The framework uses this information to predict the state of the robot after it completes its current chunk of actions. It uses that estimation to plan the next motions.

“In this way, we give the robot awareness of its future state,” Tang says.

Augmenting acceleration

On its own, this technique speeds up the robot’s motions by eliminating lag time that usually occurs between action chunks, accelerating reaction speeds more than 30-fold.  But to make their approach even faster, the MIT researchers generate coarser chunks of actions, so the robot executes a few larger steps that follow the same trajectory. This technique is called action quantisation. 

While action quantisation led to a slight dip in accuracy, it enables a robot to complete the overall task two to three times faster. However, the researchers found that simply feeding future robot states to the VLA during deployment is not enough to enable accurate and stable control of the robot. 

They developed a training-augmentation method that groups training data in such a way that the VLA learns to use future state information instead of current observations. By reusing some training data, this fine-tuning method accelerated training fivefold with no additional computational overhead. 

“Even though there is a very large model working in the background, VLASH lets the robot react and execute its actions very fast, much more like a human would. This could help to make robots for all sorts of dynamic tasks more effective,” Tang says.

When compared with baseline methods in simulation, VLASH consistently performed faster while maintaining the accuracy of robotic maneuvers. The system also outpaced these methods on real hardware in pick-and-place, stacking, and sorting tasks.

For instance, VLASH placed cubes in a box while sorting them by color twice as fast as these methods, while achieving the same 90 percent accuracy as the best baseline. The system can also perform highly dynamic tasks like playing ping-pong and whack-a-mole.

In the future, the researchers want to combine VLASH with more powerful generative AI systems called world models that can predict the robot’s actual environmental observations, in an effort to boost performance and open new applications.

This work is supported, in part, by the MIT-IBM Computing Research Lab, Amazon, the National Science Foundation, and Nvidia.

https://web.mit.edu/