The project achieved the goal of developing the central MEMS filter with new innovations in the newly developed mirror coating that significantly increases the sensitivity of the sensor. A sensor system was built that provided successful classification of ice, wet and dry surfaces under controlled conditions, but due to delays in the project it was not possible to test the system in the field. More about this is described below in connection with the review of the various work packages.
A brief summary of milestones in the H1-H5 work packages in the MOROS project:
H1) Planning and specification: The work to obtain specifications for the MEMS filter and sensor system was led by Comlight AS and was completed in 2020. The specification document has been revised as needed during the work on the development of the sensor system.
H2) Electronics and packaging: Electronics and packaging were developed to install new IR filters for testing in the relevant environment. The work on control electronics and packaging for the MEMS filter was led by Tunable AS. Tunable created a miniaturized card with microcontrollers that are considerably smaller than the previous card. Tunable has also further developed the detection functionality of the MEMS filter: now the MEMS filter can be moved in a way that provides continuous spectra considerably faster than before.
H3) Road Monitoring Sensor Design and Development: Lab setups were built based on optical simulations and designs, and have been used to distinguish between dry, wet, and icy roofing shingles under controlled conditions (see H5 demonstrator and evaluation): An area is illuminated and the collected light from that area is analyzed using the IR filter. The analysis shows the intensity for different wavelengths, and from this information is used to reveal whether the surface is dry, wet, ice-covered or snow-covered.
H4) MEMS: Newly developed IR filter chips were manufactured, but after partial significant delay due to challenges of long downtime of processing equipment. The IR filters consist of two parts that are processed separately: top chips and bottom chips. In the MOROS project, these IR filters get a new optical coating that provides significantly better transmission in the NIR range. A lot of work has gone into process development to be able to produce IR filters with this coating. The bottom pieces were completed in May 2022 while the top pieces were completed in October 2022. To ensure progress in the project, previous versions of the IR filter (without the new optical coating, developed in previous projects) were therefore used in connection with testing of the optical system in H3. Later, the new finished bottom chips (with new optical coating) were combined with old top chips (without new optical coating) to realize about half the gain of the new coating. This "hybrid" solution was used for testing in the H5 Demonstrator and evaluation, discussed below. Finally, the final IR filter was made, consisting of bottom chips and top chips with new mirror coatings on them.
Another goal of the project was to demonstrate the possibility of assembling MEMS fitlers at disc level. This work was initiated, but due to equipment downtime towards the end of the project, this work was not concluded.
H5) Demonstrator and evaluation: Demo testing and data collection was conducted by Comlight. The optical system developed in H3 was used to distinguish between dry, wet, ice-covered and snow-coated roof shingles under controlled conditions. In the first phase, a freezer was used and measurements were carried out indoors at a distance of 1.8m. In the next phase, the experiment was moved out to a garage where you could choose between taking measurements with and without daylight when opening the garage door. Then measurements were carried out at a distance of 3.6m, and a peltier element cooled plate was used to carry out measurements while the surface went from wet to ice. Then snow was taken in from outside and put on the plate. Despite the fact that the "hybrid" IR filter (described above in H4) was used (which has a slightly lower performance than the desired filter), successful measurements were obtained that provide clear classification between dry, wet, ice and snow using various algorithms (Principle Component Analysis and Linear Discriminant Analysis). An example is shown below ("icb" stands for ice-coated and "snb" stands for snow-covered): The four categories are clearly separated from each other, and a machine learning algorithm will therefore be able to easily distinguish between the different states.
Prosjektet nådde målet om å utvikle det sentrale MEMS-filteret med nye innovasjoner i nyutviklede speilbelegget som øker sensitiviteten til sensoren betraktelig. Det ble bygget et sensorsystem som ga vellykket klassifisering av is, våt og tørre overflater under kontrollerte forhold med relativt robust klassifisering.
Resultatet gir grunnlag for videre utvikling av løsningen som kan være med å forbedre varsling av glatte veistrekninger og raskere respons på strøing og/eller varsling av billister.
I dette prosjektet vil Comlight og Tunable sammen med SINTEF utvikle et sensorsystem for overvåkning av veioverflaten (is, snø og vann) som er integrert i fast infrastruktur langs veiene. Systemet muliggjør varsling til bilister om veiforholdene fremover, og vil gi verdifull informasjon til veivedlikehold og behov for brøyting, strøing og salting.
Det er flere selskaper som tilbyr optisk deteksjon av veibanen, men disse er både store og for kostbare. Comlight vil utvikle et instrument som er rimelig i pris og enkelt å installere slik at storskala utrulling av veiovervåkningssensorer er mulig. Dette vil de oppnå ved å benytte et mikroelektromekanisk system (MEMS) utviklet ved Sintef og kommersialisert for gassdeteksjon av Tunable. Selve MEMS-brikken fungerer som et miniatyrisert optisk spektrometer som kan transmittere lys innenfor et kontinuerlig spektrum av infrarøde bølgelengder. Gjennom prosjektet vil Tunable kunne etablere volumproduksjon av filtermoduler og selge slike utover eget bruk, inkludert til is-deteksjon.
Sentrale FoU utfordringer som må løses er følgende
- Framstille MEMS filtre for riktig spektralt område for å separere is, vann, snø og tørre overflater
- Utvikle en kompakt og rimelig optisk pakke med styringselektronikk for MEMS-filteret
- Demonstrere prosesser som kan gjøre masseproduksjon av MEMS filtre mulig
- Utvikle en kompakt demonstrator som er enkel å installere og kan klassifisere veioverflaten
Om prosjektet lykkes vil det være mulig å oppnå tilstrekkelig tetthet av sensorer gjennom integrasjon i fast infrastruktur langs veiene (lyktestolper og lysarmaturer) til å få full oversikt over vanskelige veiforhold og vil kunne muliggjøre en behovsstyrt brøyting, strøing og salting av vei og sanntidsinformasjon om veiforhold til bilister. Et vellykket prosjekt sum utvikler teknologi for rimelige justerbare optiske filtre vil også åpne muligheter innen en rekke andre markedsområder som materialgjenvinning og hyperspektral avbildning.
Funding scheme:
NANO2021-Nanoteknologi, nanovitenskap, mikroteknologi og avanserte materialer