Planet of Mimics (PoM) is my (Zhen Hua, Zhou) personal website where I document publicly shareable research and projects. This site is intended for Siemens colleagues as well as external friends, clients, and partners, and for anyone interested in my work. Most content consists of academic research and software prototype demonstrations. Any information related to my employer (Siemens AG) has been de-identified to ensure compliance and security.

The Name

The term "mimics" actually comes from American author Michael Crichton's science fiction novel Prey, which describes a type of nanorobot. A swarm of self-evolving nanobots with collective intelligence breaks free from laboratory control, becomes a deadly predatory threat, and can imitate and fully transform into any object.

The core traits embodied by mimics include:

• 1. A high degree of realism, both in static appearance and in dynamic behavior.
• 2. Adaptation and evolution based on assimilating and summarizing human technologies, rather than simple replacement.
• 3. AI-driven intelligence — effectively a form of general artificial intelligence.

I find the name both interesting and symbolic, because I have been researching and developing a technology I call "Dynamic Behavior Modeling (DBM)". Its ultimate goal is to create intelligent entities that can highly simulate and adapt to environmental changes, which in many ways resemble the mimics in the novel.

This website is like a planet (a platform) I am trying to portray and even build — a place that can host all these charming mimics and form an ecosystem. I also hope to draw special inspiration and even technical breakthroughs from it.

Language

The primary language of this website is English, chosen to reach the widest possible audience. Due to time constraints, I do not plan to support multiple languages; if needed, please use AI tools to assist with translation. English is not my native language (my native language is Mandarin Chinese), so please excuse any inaccurate expressions.

There is a serious contradiction between the rapid development of AI today and the overall technological backwardness we observe in industry. Take factories as an example: after so many years, the vast majority of factories worldwide are far from fully digitized or highly automated. Much of what is called digitalization began by purchasing many different industrial software packages and hardware controllers, yet software, hardware, and processes remain fragmented. Very few reach the level of Siemens' Amberg plant. Overall, software engineering in the OT domain lags behind IT by roughly thirty years, and this reality has not been fundamentally resolved. Given such an industrial foundation, will the arrival of AI make all problems disappear? I don't think so.

AI fully understanding IT may not be an issue in the future, but getting AI to understand — and especially to control — OT involves a huge gap. I personally do not believe many startups' current industrial approaches are correct. This depends on several factors, including:

• AI has difficulty obtaining OT data and domain knowledge needed for learning.
• AI's randomness conflicts with OT's requirements for precision, real-time behavior, and strict safety.
• The OT world is a large, fragmented collection with severe lack of standardization; even if AI were deployed, interoperability and connectivity would be hard to achieve. Monolithic, closed giant software systems are also widespread.

All of my research — including the studies, products, and solutions documented on this website — follows the same mission: to build a standardized, abstract semantic model that both humans and AI can understand, adapts to IT and OT, and describes the virtual and physical worlds within the same data space. Based on this foundation, I develop the corresponding engineering tools, runtimes, and agents that apply best user‑experience practices for humans and provide holographic, learnable data for AI, ultimately achieving true digitalization.

• 1. Simplicity · Fully Modular Design
  
  Simplicity is not only visual clarity but also refinement in functionality and structure. It emphasizes removing redundancy and highlighting the core, allowing users to understand and use the system in the shortest possible time. Fully modular design means the system is composed of multiple independent and replaceable functional units; these modules have clear interfaces and minimal dependencies, and can be freely combined, swapped, and upgraded like building blocks. This philosophy reduces development and maintenance costs, adapts flexibly to different scenarios, and extends product lifecycle.
• 2. Semantic Design · AI-Friendly
  
  The core of semantic design is to give data, interfaces, and interactions clear, structured meaning rather than relying solely on form or position to interpret content. With a good semantic structure, AI can parse, learn, and reason more accurately, enabling more natural human-computer interaction. This not only makes it easier for AI to understand user intent but also smooths future intelligent extensions. For example, semantic data labels, interface descriptions, and behavior definitions can be used directly as inputs for machine learning and automated decision-making, reducing intermediate conversion loss.
• 3. Standardization · Building an Ecosystem
  
  The purpose of standardization is to establish unified interfaces, protocols, naming conventions, and quality requirements so that products from different teams and vendors can collaborate seamlessly. Through standardization, interoperability between products is enhanced, forming an open and sustainably extensible ecosystem. Such an ecosystem not only attracts more developers and partners but also provides users with a consistent experience and service within the same framework, thereby enhancing the system's competitiveness and vitality.