auto·botik

auto - automated | botik - robotics

For over two decades, Autobotik has worked at the edge of automated systems and robotics - engineering automated driving technology and power electronics across Europe, Asia, and the US, and now applying that same engineering discipline to the sensor fusion, perception, and safety-verification methods that physical AI systems require. Resembler.ai, the platform built on Autobotik's safety-verification methodology, has been licensed by CETRAN - Singapore's automated-vehicle testing centre, owned by the Land Transport Authority (LTA) - for more than two years, surfacing edge cases in the validation of automated driving systems. The throughline is simple: build systems that act in the physical world, and prove they can be trusted to.

2018 ITS - Symantic Gap and Heuristic Gap
in Self Driving AI systems

Most of today's AI conversation is about text and digital data - language models trained on words, reasoning about the world through what's written about it. Physical AI is different: it has to act in the real world itself, where the rules aren't written down anywhere. A child knows a ball bounces and a bottle breaks long before they can read a word about gravity or material science - not from a book, but from throwing things, dropping things, hearing the difference between a bounce and a break, feeling their own balance shift when they stumble, and the reaction of a parent the one time they got it wrong. None of that comes from a page; it comes from being in the world and dealing with what happens next.

Physical AI systems have to build that same understanding from perception, action, and consequence - not text. A drone that misjudges the wind doesn't get to revise its answer; it acts, and the physical world responds immediately. Autonomous systems are physical AI: automated driving, drones, and robotics are physical AI systems that move through the world instead of sitting still.

Almost every new physical AI system depends on power electronics - the motors, inverters, and converters that turn a decision into motion. That's a harder integration problem than it looks: the actuation layer has to be designed for functional safety, not just performance, with torque and current sensors integrated into closed-loop feedback across the full system - so it knows not just what it commanded, but what the motor actually delivered.

Physical AI does not read
the manual

Where simulation runs out

Traditional systems engineering verifies against the V-model: define requirements up front, then test each stage against that fixed specification. It works when correct behavior can be fully specified in advance - but physical AI can't be specified that completely, and three gaps show why. There's a semantic gap between what a requirement says and what a real situation actually means, because the real world doesn't arrive as a clean, described scenario. There's a heuristic gap: test cases are built from edge cases engineers already know to look for, so they catch known failure modes and miss the ones nobody thought to write down. And there's the incompleteness of synthetic data - simulation can generate enormous volumes of scenarios, but only the ones it was designed to generate, which means it reflects the limits of its creators' imagination rather than the real distribution of what a system will actually encounter.

Resembler.ai, the platform built on Autobotik's safety-verification methodology, closes these gaps by verifying AI with AI - using large language models to evaluate system behavior against a real-world baseline of over 500 million kilometers of recorded human driving, rather than a fixed scenario library. CETRAN, Singapore's automated-vehicle testing centre under the Land Transport Authority, has licensed it for more than two years to surface exactly the edge cases scripted testing misses.

www.resembler.ai

Safe Deployment of Autonomous Technology

Ensuring the safety and reliability of self-driving mobility and intelligent infrastructure solutions is critical as autonomous technologies move toward deployment in public, private, and industrial domains. While traditional functional safety standards such as ISO 26262 provide a foundation for system reliability, they are insufficient for addressing the unique challenges of AI-driven autonomy.

Autobotik actively supports safety case development for autonomous mobility and infrastructure, going beyond functional safety to tackle key gaps in Safety of the Intended Functionality (SOTIF – ISO 21448), UL 4600, and real-world risk assessment methodologies. Our approach ensures that automated driving systems (ADS) and mobility infrastructure are not only functionally safe but also resilient to real-world uncertainties and edge cases.

Functional safety under ISO 26262 ensures system reliability by addressing hardware failures but falls short in handling AI-driven perception, decision-making, and real-world unpredictability. SOTIF (ISO 21448) expands safety to cover unknown risks, sensor limitations, yet open-world scenarios and edge cases require further validation. UL 4600 provides a structured safety assurance framework but still lacks real-world testing, human factor integration, and continuous monitoring. To address these challenges, Autobotik offers an AI platform Resembler.ai, integrating advanced risk assessment, real-world validation, and AI-powered safety analytics. This partnership enables a rigorous approach to verifying autonomous system behavior, ensuring self-driving mobility and infrastructure solutions meet and exceed evolving safety standards. More details at www.resembler.ai.

2015 SF to NYC Coast to Coast Automated Drive

Navigating the New Geopolitics of Automotive Supply Chains

Leveraging our deep expertise in global automotive technology and supply chain integration, Autobotik helps companies adapt to shifting trade policies, regionalized production strategies, and evolving technology ecosystems. With a long-term understanding of how automotive innovation can be distributed across global networks, we support OEMs, Tier 1 suppliers, technology firms, and policymakers in building resilient, future-proof strategies for electrification, autonomy, and connected mobility.

Autobotik has been at the forefront of these shifts, actively supporting industry players and consultants since our Booz Allen Strategy+Business feature “Making Offshore Engineering Pay Off - How some companies send design work overseas without fear of diminished quality or intellectual property theft”.

Our deep understanding of the Software-Defined Vehicle (SDV) transformation, domain-specific technology development, and global supply chain dynamics allows us to guide OEMs, Tier 1s, startups, and institutions in rethinking where critical innovations should take place and how to build a resilient and future-proof industry strategy.

Shaping Future Sustainability: Electrification, Autonomy and Smart Cities

The convergence of four mega-trends electrification, shared economy, connectivity and autonomous driving is changing the way we think about mobility, transportation, logistics and smart city robotics. Although each trend is profoundly disruptive on its own, the combination of all four is creating new challenges and opportunities.

Today, the automotive industry is dominated by traditional OEMs, which design, manufacture, and sell vehicles through capital-intensive mass production and extensive distribution networks. However, as electrification, autonomy, connectivity, and the shared economy converge, the future of mobility will be shaped by new players, innovative business models, and technology-driven ecosystems. The shift toward sustainable, on-demand, and intelligent transportation solutions is creating opportunities for startups, technology companies, cities, and institutions to play a pivotal role in reshaping mobility, logistics, and smart city infrastructure.

Autobotik offers a range of specialist services to startups, the automotive industry, cities, institutions, and organizations or those wishing to implement and/or prepare for new mobility solutions. We advise and help implement in the domains of Autonomous Mobility Solutions, Automated Driving, Vehicle Electrification and the future Smart City Digital Eco-System, on technology, products, required systems, technology, automotive business, eco-system requirements, business models, executive training and go-to-market strategies.

Electrification Technology

We offer support in Power Electronics (PE) design and vehicle systems integration for EV, HEV and new smart city mobility applications. Our footprint in Asia and Europe provides the necessary support and capabilities for technical, vehicle architecture, and global sourcing support from selection to management and quality assurance of components, sub-systems and electrification systems. A few things we have done in this area:

Strategy and Technology development support for EV development
Technology guidance during new EV/HEV product development

Our Team

We are a global team of highly-motivated individuals from diverse cultural backgrounds, usually working together in the Singapore-based office and sometimes operating together from various locations around the world. The team has extensive experience in all domains required to implement new urban automated mobility and many years of executive automotive experience implementing automated driving technologies, pilots and automotive electrification in the US, Europe and Asia.

Strategic development of organization capability

Product Engineering in today’s world is a global activity with fast changing technology and geo-political transformation. In light of this, a large part of footprint decision is driven by cost, talent availability and coordinated technology development.

With our decades of experience in managing global product engineering and successfully delivering local and global programs, we offer a holistic assessment and recommendation framework for development of organization capability. As part of this, we also consider supply chain, manufacturing and product integration complexity. This work is based on <booz allen….>. Here’s what we do

Technology Roadmap Development and Global Engineering Consulting
Enterprise Maturity Assessment in Process, Technology, Capability and Products
Footprint plan and CoE (Centers of Excellence) strategy for optimum talent utilization
IP Strategy