“Multimodal design must follow a less-is-more philosophy”

Hanna Lukashevich has been with Fraunhofer IDMT for more than eighteen years and has led its Semantic Music Technologies department for over eleven of them. Her work spans AI-based vocal detection, musical meta-feature extraction and the transfer of semantic music-analysis methods to industrial audio — demonstrating how techniques developed for music can be applied effectively to automotive use cases.

Automotive Logistics & Supply Chain Europe

17-19 March | Bonn, Germany

At the intersection of sound, lighting and intelligent cabin sensing, she offers key insights into the next generation of multimodal in-vehicle experiences. Building on this expertise, we spoke with the scientist about the technological demands shaping future audio-reactive interior lighting.

ADT: We are in the midst of a dynamic and disruptive decade for the automotive industry. From your perspective, what are the biggest challenges the interior lighting sector will face over the next five years?

Lukashevich: A key trend is clear: simplification. Less is more will continue to shape interior lighting, driven by strong cost pressure and the need to manage increasing technical complexity. At the same time, safety considerations will gain importance. The debate around distraction by light will grow, meaning that intensity, motion and dwell times will require conservative and well-defined limits. From an audio-technology perspective, an additional set of challenges becomes crucial when lighting reacts to sound. The most difficult aspects over the next five years lie exactly at the interface between audio and light. First, achieving low latency and tight synchronization is essential. For beat- or voice-driven cues to feel natural, end-to-end detection to LED response must reliably stay within a few tens of milliseconds, even as the number of addressable LEDs increases. Second, designing balanced visualizations across audio genres is becoming more demanding. Lighting that simply follows bass works for certain music styles but fails for speech, classical or ambient content. Future systems must react to richer audio features such as tempo, loudness, timbre and mood, and adapt color palettes and motion accordingly. Third, finding the right trade-off between responsiveness and stability is critical. Over-sensitive mappings cause flicker during rapid audio changes, while heavily smoothed ones lose rhythm and emotional impact. Solutions lie in adaptive smoothing, tempo-grid alignment and guardrails for color and motion transitions.

“IMSE enables seamless HMIs like hidden-until-lit controls”

Integrating light and electronics directly into structural components marks a shift in how car interiors are designed. In this interview, Hubert Bieder of Tactotek explains how IMSE (Injection Molded Structural Electronics) enables functional surfaces that merge lighting, sensing, and control.

Benjamin Müller

What additional considerations come into play?

Beyond audio-reactive lighting, interior lighting must be treated as a core design element early in the development process. Scaling up introduces its own technical demands: per-LED calibration and thermal compensation to maintain uniform whites and hues; low-noise electronics to prevent dense LED buses from interfering with microphones or audio DSP; power-aware animations that respect EV energy budgets; privacy-by-design for any in-cabin audio analytics; and forward compatibility with emerging domain or Ethernet lighting architectures. With today’s protocol fragmentation, momentum toward standardization can also be expected, both in communication protocols and shared application-level semantics, to ensure interoperable and safe behavior.

Your research covers audio signal processing, machine learning and semantic technologies. How can insights from this work inspire innovation in automotive interiors, particularly in creating multimodal user experiences?

Audio signal processing, machine learning and semantic technologies enable a shift from simple reactive color effects to truly authored multimodal in-cabin experiences. Real-time music analysis, capturing beat, tempo, loudness, timbre, mood and speech-music transitions, allows lighting to respond in a way that enhances the atmosphere without distracting the driver. Object-based audio adds spatial context; by linking individual sound objects to specific lighting zones, vocals, drums or effects can gain subtle visual counterparts that increase immersion and help passengers perceive the scene more clearly. A semantic layer, combining information such as time, route context or user preferences, turns these features into coherent lighting scenes with consistent choices for color, motion and intensity that remain robust across genres. To push quality further while keeping latency low, a hybrid pipeline is promising. Pairing real-time control with AI-driven offline pre-authoring enables the generation of visualizations in advance, which can be stored as lighting descriptors and retrieved via audio fingerprinting when a known track plays. This delivers reliable and high-quality behavior across different music styles while avoiding typical issues such as oversensitivity or sluggish smoothing. Throughout all of this, multimodal design must follow a less-is-more philosophy: control complexity and cost, and always prioritize safety by avoiding rapid contrast changes or patterns that could distract the driver.

“Higher costs require more flexibility in procurement”

Interior lighting in cars is becoming increasingly intelligent, connected, and personalised. Otmar Rauchensteiner from Dräxlmaier explains how software and AI are shaping the lighting architecture of the future - and how global trade conflicts affect it.

Benjamin Müller

Fraunhofer IDMT is working on anomaly detection and audio-visual content verification. How could such technologies contribute to smarter interior systems, for example by linking sound, sensing and adaptive lighting to improve safety and comfort?

Fraunhofer IDMT’s work on anomaly detection and audio-visual content verification can make interior systems smarter by linking sound, sensing and adaptive lighting to enhance both safety and comfort. Environmental sound analysis allows the system to detect and localize hazards such as sirens, horns or approaching two-wheelers and to respond within tens of milliseconds. For example, the vehicle can briefly attenuate entertainment audio and guide attention with subtle, zone-specific light cues on the side of the detected event, ensuring that the driver notices the hazard without being distracted. The same tools can monitor mechanical noises inside the vehicle. Wear-related sounds such as rattling, squeaking, banging, bearing whine or pump cavitation can be transformed into calm, localized maintenance prompts instead of intrusive alerts, helping prevent failures and reducing complaints related to noise, vibration and harshness. Audio-visual content verification further enhances system trust. It can detect spoofed or synthetic voice commands; if confidence is low, the system can block the action, request a simple confirmation and indicate this state with a distinct verification needed light pattern. All processing should prioritize on-device execution using minimal, short-lived features to protect privacy. Lighting responses must remain clear, consistent and non-overloading to ensure safety and comfort at all times.

When the car picks up the bill

As mobility becomes more digital, tokenisation offers new ways to manage payments, component histories and vehicle identities. German manufacturers are now testing blockchain to bring these concepts into real operations.

ADT editorial team