Statistics Fractiles Transportation Engineering Traffic Accident Analysis

Measure of Natural Tendency  Range Design of Horizontal Curve  Accident rate for 100 million
= 𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 − 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 vehicles per miles of travel in a
 Mean, x̅, μ → average
 Minimum radius of curvature segment of a highway:
→ Mode Stat 1-var  Coefficient of Range 2
→ Shift Mode ▼s Stat Frequency? on 𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 − 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 v A (100,000,000)
= R= R=
→ Input 𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 + 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑑𝑎𝑡𝑢𝑚 g(e + f) ADT ∙ N ∙ 365 ∙ L
→ AC Shift 1 var x̅ R → minimum radius of curvature
A → no. of accidents during period of analysis
 Quartiles e → superelevation
f → coeff. of side friction or ADT → average daily traffic
 Median, Me → middle no. when n is even
skid resistance N → time period in years
n+1 1 2 3 v → design speed in m/s L → length of segment in miles
Me th = Q1 = n Q2 = n Q3 = n g → 9.82 m/s2
2 4 4 4
 Accident rate per million entering
1 n n when n is odd  Centrifugal ratio or impact factor
Me th
= [( ) + ( + 1)] vehicles in an intersection:
2 2 2 1 1 1 2
Q1 = (n + 1) ; Q1 = (n + 1) ; Q1 = (n + 1) v A (1,000,000)
4 4 4 Impact factor = R=
 Mode, Mo → most frequent gR
 Interquartile Range, IQR
ADT ∙ N ∙ 365
R → minimum radius of curvature
Standard Deviation v → design speed in m/s A → no. of accidents during period of analysis
= 𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒 − 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒
g → 9.82 m/s2 ADT → average daily traffic entering all legs
 Population standard deviation = Q3 − Q1 N → time period in years
→ Mode Stat 1-var Power to move a vehicle
 Coefficient of IQR  Severity ratio, SR:
→ Shift Mode ▼ Stat Frequency? on P = vR
𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒 − 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒
→ Input = P → power needed to move vehicle in watts f∙i
𝑙𝑎𝑟𝑔𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒 + 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒 SR =
→ AC Shift 1 var σx v → velocity of vehicle in m/s
Q − Q1 R → sum of diff. resistances in N
f∙i∙p
= 3
 Sample standard deviation
Q3 + Q1 f → fatal
Design of Pavement i → injury
→ Mode Stat 1-var  Quartile Deviation (semi-IQR) = IQR/2 p → property damage
→ Shift Mode ▼ Stat Frequency? on  Rigid pavement without dowels
→ Input  Outlier  Spacing mean speed, US:
→ AC Shift 1 var sx
→ extremely high or low data higher than 3W ∑d n
or lower than the following limits: t=√ Us = =
f ∑t ∑ 1
NOTE:
Q1 − 1.5IQR > x ( )
If not specified whether population/sample U1
in a given problem, look for POPULATION. Q 3 + 1.5IQR < x  Rigid pavement with dowels
Coefficient of Linear Correlation  Time mean speed, Ut:
 Decile or Percentile 3W 3W
or Pearson’s r d
m t=√ t=√ ∑
im = (n) 2f 4f ∑ U1
→ Mode Stat A+Bx
10 or 100 Ut = t =
→ Input (at the edge) (at the center) n n
→ AC Shift 1 Reg r t → thickness of pavement Ʃd → sum of distance traveled by all vehicles
Normal Distribution W → wheel load Ʃt → sum of time traveled by all vehicles
NOTE: f → allow tensile stress of concrete Ʃu1 → sum of all spot speed
-1 ≤ r ≤ +1; otherwise erroneous 1/Ʃu1 → reciprocal of sum of all spot speed
 Flexible pavement n → no. of vehicles
 Population standard deviation
W  Rate of flow:
Variance t=√ −r
 Z-score or
standard score → Mode Stat 𝜋f1 q = kUs
 standard deviation = σ
or variate → AC Shift 1 Distr f1 → allow bearing pressure of subgrade q → rate of flow in vehicles/hour
 variance = σ2 r → radius of circular area of contact
left of z → P( k → density in vehicles/km
x−μ between wheel load & pavement
 relative variability = σ/x z= uS → space mean speed in kph
σ right of z → R(
Mean/Average Deviation bet. z & axis → Q(  Thickness of pavement in terms  Minimum time headway (hrs)
x → no. of observations
μ → mean value, x̅ → Input of expansion pressure = 1/q
 Mean/average value σ → standard deviation expansion pressure
t=  Spacing of vehicles (km)
b pavement density
1 = 1/k
mv = ∫ f(x)dx Exponential Distribution
b−a a  Stiffness factor of pavement
 Peak hour factor (PHF)
P(x ≥ a) = e−λa = q/qmax
 Mean value
P(x ≤ a) = 1 − e−λa
Es 3 s

SF = √
1 b P(a ≤ x ≤ b) = e−λa − e−λb Ep
RMS = √ ∫ f(x)2 dx ES → modulus of elasticity of subgrade
b−a a EP→ modulus of elasticity of pavement

Discrete Probability Distributions Walli’s Formula

π
 Binomial Probability Distribution 2 [(m − 1)(m − 3)(m − 5) … (1 or 2)][(n − 1)(n − 3)(n − 5) … (1 or 2)]
P(x) = C(n, x) p q x n−x ∫ cosm θ sinn θ dθ = ∙α
0 (m + n)(m + n − 2)(m + n − 4) … (1 or 2)
where:
p → success NOTE:
q → failure
α = π/2 for m and n are both even
 Geometric Probability Distribution α =1 otherwise
x−1
P(x) = p(q ) Tip to remember:
Fibonacci Numbers
 Poisson Probability Distribution 𝑥2 − 𝑥 − 1 = 0
n n
x −μ 1 1 + √5 1 − √5 Mode Eqn 5
μ e
P(x) = an = [( ) −( ) ]
x! √5 2 2 𝑥=
1 ± √5
2

Period, Amplitude & Frequency

x = r cos θ
Period (T) → interval over which the graph of y = r sin θ
function repeats r = x2 + y2
Amplitude (A) → greatest distance of any point y
on the graph from a horizontal line which passes θ = tan−1
halfway between the maximum & minimum
x
values of the function
Frequency (ω) → no. of repetitions/cycles per unit
of time or 1/T

Function Period Amplitude

y = A sin (Bx + C) 2π/B A
y = A cos (Bx + C) 2π/B A
y = A tan (Bx + C) π/B A