Table 2. The solution streams: get automatic detection of low velocities and embed a new procedure (TARP) that allows Geotechnical Engineers to know when a movement in the lower range of velocities is potentially concerning. Automatic Detection of Low Velocities AI Early Warning Pattern Detection Working with OEM, it has been showing that the SMA method is a valid means of calculation lower velocities Clustering methods used to group similar movement patterns for the same rockmass materials and failure mechanism Software update with OEM enables automatic calculation of low velocity movements Clustering highlighted common displacement rate behaviour pre-failure, during failure, post failure and stable behaviour No hardware updates required to existing OEM Ground-Based Radars Data driven approach shows when lower velocities signal risk and can inform key decision points in a new TARP The same data is used as the current real-time processes This paper does not seek to duplicate detailed descriptions of radar theory, interferometric processing, or algorithmic optimisation, which are addressed in companion technical publications. Instead, the emphasis is placed on how analytical capability and engineering expertise were combined to manage uncertainty, support interpretation, and enable scalable early warning capability. This framing reflects the view that technical performance and organisational design are inseparable in safety critical systems. 3. METHODOLOGY: TRANSDISCIPLINARY OPERATING MODEL AND DELIVERY The methodology adopted in this work treats collaboration as a deliberately engineered system rather than an informal by-product of project delivery. A central element of this operating model was the co-location of geotechnical engineers, data scientists, and technology specialists, which enabled real-time dialogue, rapid clarification of assumptions, and continuous alignment of analytical outputs with geotechnical meaning. , while each discipline retained ownership of its decisions—while still enabling frequent interaction at key decision points, avoiding the fragmentation that typically occurs in linear handovers. This environment supported iterative cycles of joint problem definition, analytical exploration, expert interpretation, and refinement, consistent with evidence showing that iterative co-creation is essential for complex problem-solving under uncertainty. Research on transdisciplinary practice further highlights that effective integration depends on collaboration culture, shared understanding, and active engagement between disciplines—conditions. A defining feature of the team was the presence of multiple boundary-spanning individuals who developed fluency across domains, enabling them to translate analytical constructs into engineering insights and express operational concerns in analytically tractable ways. By distributing boundary-spanning capability
RkJQdWJsaXNoZXIy MTM0Mzk2