運用自製低頻噪音(< 400 hz)與CO₂ 監測器進行實測，獲知汽車排氣噪音明顯隨平均車速增加而降低，遞減率為10.21 mV•hr/km，且不易受風場影響，可準確反映車行狀況，並具備低成本、低耗能與易於維護，是非常理想監測器，而最佳裝設位置為白線後8 m。CO₂ 則易受風場影響而明顯混合，濃度隨平均車速的變化小，遞減率為1.58 ppm•hr/km，較不適用。結果顯示820PCU/hr-lane 是交通壅塞臨界點，此點車速為25 km/hr。成功製備「即時監測−反應之交通號誌控制系統」，證實引擎排氣噪音可反饋做為即時交通流量與車輛排放的控制。最後，模擬交通改善後狀況，計算出避開壅塞臨界點需增加綠燈秒數，在本研究中最佳增加值約4 sec/cycle。並推估台中市五條主要幹道尖峰時NOx、CO與THC 分別削減2,768、148,027 與41,373 kg/year，顯現採用引擎排氣噪音反饋至道路交通流量與車輛排放控制的成效，相當好。最後，提出一個控制城市交通號誌的簡易方法。
The relations between vehicle average speed with low frequency engine exhaust noise(<400 Hz) and CO₂concentration were investigated with self-prepared noise and CO₂ monitors on-road. After field measurements, I found that the real relation of the vehicle speed with noise intensity was more obvious than with CO₂ concentration, the R-squares were 0.632 and 0.152, respectively. The possible reason is CO₂ in the field mixes with air flux very easily, but low frequency noise was measured with high precision and considerable speed. The lapse rates of CO₂ and low frequency noise are 1.58 ppm•hr/km and 10.21 mV•hr/km, respectively. The optimum measuring post of a noise monitor is about 8 m before the stop line. The critical point for a traffic jam is 820 PCU/hr-lane with a speed of 25 km/hr. Increasing the time of green light operation for avoiding this critical point was evaluated. In this study optimal increase was about 4 sec/cycle. Finally, the success of low frequency noise sensors as a real-time vehicle monitor was shown, and a real-time monitoring-response traffic signal control system was set up. The reduction amounts of NOx, CO, and THC from the vehicles emission during rush hours on five main roads in Taichung city were 2,768, 148,027, and 41,373 kg/year, respectively. The results indicate that engine exhaust noise feedback to traffic flow and vehicle emission control on-road is useful and accurate. Finally, an easy method for urban traffic light control was suggested.