PENGEMBANGAN METODE PREDIKSI PROBABILITAS KECELAKAAN PADA SUATU RUAS JALAN ANTARKOTA BERDASARKAN KONDISI GEOMETRIK JALAN DAN KARAKTERISTIK LALU LINTAS

Abstrak

Abstract

Accident data is very important to figure out the number of accidents that occur on the road segment in some periods of time. If accident data is not available, another method is needed to determine the frequency of accidents, with the frequency of accidents replaced by a prediction of accident probability. Based on the literatures, it is found that the predictions of accident probabilities can be done based on the geometric conditions of the road and traffic characteristics. The probability values are obtained from the calculation of the percentage of the total score on the location of the road. The score is given based on the magnitude of the existing condition deficiency against the specified standard or criteria, with a score of 1 for conditions that are in accordance with the standard, while a score of 5 for the conditions that are not in accordance with the standards. Predictor parameters for 2 lane 2 ways undivided rural road are lane width (m), lane widening on curve, shoulder width, shoulder type, stopping sight distance, passing sight distance, radius of curve, maximum superelevation, tangent length between 2 curves, the present of transition curve, curve radius ratio of 2 adjacent curves, maximum length of tangent, grade (%), maximum length of grade, side slope, clear zones, hazards type, lighting, signings and markings, driveways density (number/km), proportion of heavy vehicles (%), volume of pedestrians (person/day), special alignment combinations and the differences between operating speed (V₈₅) and speed limit (V_limit). From this method would be generated 5 accident probabilities categories which are “very small”, “small”, “medium”, “big” and “very big”.

Keywords: traffic accident, accidents frequency, accident probability, road geometric

Abstrak

Data kecelakaan sangat penting untuk mengetahui jumlah kecelakaan yang terjadi pada segmen jalan untuk jangka waktu tertentu. Jika data kecelakaan tidak tersedia, diperlukan metode lain untuk mengetahui frekuensi kecelakaan, dengan frekuensi kecelakaan digantikan dengan perkiraan probabilitas kecelakaan. Berdasarkan literatur diperoleh bahwa prediksi probabilitas kecelakaan dapat dilakukan berdasarkan kondisi geometrik jalan dan karakteristik lalu lintas. Nilai probabilitas diperoleh dari hasil perhitungan persentase total skor pada lokasi ruas jalan. Pemberian skor dilakukan berdasarkan besarnya defisiensi kondisi eksisting terhadap standar atau kriteria yang ditentukan, dengan skor 1 untuk kondisi yang sesuai standar, sedangkan skor 5 untuk kondisi yang paling tidak sesuai standar. Parameter predictor adalah lebar lajur, pelebaran lajur pada tikungan, lebar bahu, tipe bahu, jarak pandang henti, jarak pandang mendahului, jari-jari tikungan, super-elevasi maksimum, panjang tangent di antara 2 tikungan, keberadaan lengkung peralihan, rasio jari-jari 2 tikungan berdekatan, panjang tangent maksimum, kelandaian (%), panjang landai, side slope, clear zone (m), jenis hazard, penerangan jalan, rambu dan marka jalan, driveways density (jumlah/km), proporsi kendaraan berat (%), volume pejalan kaki (orang/hari), kombinasi alinyemen khusus, serta selisih kecepatan operasional (V₈₅)dari V_limit. Dihasilkan 5 katagori probabilitas, yaitu “sangat kecil”, “kecil”, “sedang”, “besar”, dan “sangat besar”.

Kata-kata kunci: kecelakaan lalu lintas, frekuensi kecelakaan, probabilitas kecelakaan, geometrik jalan