The safety level of an interchange is better than the safety level of an intersection for similar traffic volumes, while the intersection still operates well without delays. An intersection can have a potential of traffic collisions, specifically side-head or side/angle collisions.

An intersection might be converted to an interchange in order to alleviate safety problems related to high speed on the intersecting highways (specifically before isolated intersections), insufficient gaps, sight distance restrictions, difficult turns, and restrictions of vertical and horizontal alignments near the intersection. When the intersection experiences severe accidents, the first stage for improvements is the implementation of local and focused geometric solutions with minimal cost. If these solutions are not useful for improving traffic safety, then grade separation or a compact grade-separated junction (CGSJ) or an interchange might be justified.

The vital elements of interchange safety refer to interchange configurations, traffic controls and spacing. Safety research generally focused on most common interchange types: diamonds and cloverleafs [4]. Within the interchange individual components safety research concentrated on ramps, ramp terminals, speed change lanes, alignment and spacing. Ramp safety elements are acceleration and deceleration lanes, weaving sections, ramp alignment, and ramp terminals. The interchange alignment factors include sight distance which is discussed in this study, vertical and horizontal curves which depend on sight distance policy and grades.

Figure 1 presents a schematic diamond interchange with its individual components. Ramps, acceleration lanes, and crossroad ramp terminals are considered as individual elements [5].

Twomey et al. [4] concluded that it is safer to merge or diverge a given traffic volume of vehicles with or from a freeway at several minor flow ramps (CD road etc.) than at single high volume on-ramps and off-ramps.

Overall, designers can reduce drivers’ pressure at interchanges by keeping alignment simple, direct, and consistent, and by providing sight distance greater than the stopping sight distance (i.e. decision sight distance) which results in employing above minimum design criteria for interchange geometric elements [1].

The paper introduces implementation of sight distance policy for: ramps’ geometric design, ramp terminals’ components, and basic lanes’ reduction. These implementations of sight distance types (including their relevant models and parameters), for interchange design, especially on ramps, ramp terminal influence zones (along acceleration lane and prior to exit ramp along the major highway), and basic lanes’ reduction prior to or after two lane exit ramp terminals are not available in the literature.

Section 2 introduces a unique decision sight distance model to be implemented in: 1) ramp design when the ramp is connected to the minor intersecting highway by a signalized intersection with additive lanes (presented in Section 3); 2) prior to interchange ramp terminal zone along the main lanes, and along the acceleration lane merging to the highway (presented in Section 4). Surplus practical applications of decision sight distance are relevant to the design of basic lane reduction along freeways and interchanged highways. These applications are documented in Section 5. The final part of Section 2 includes practical implementations of stopping sight distance, specifically for ramp design when the ramp is connected to the minor intersecting highway by an intersection (unsignalized, roundabout, or signalized), without additive lanes (presented in Section 3).

Decision sight distance is measured from the driver’s eye height of 1.05 m to an object height of 0.6 m. It is basically implemented prior to the interchange ramp terminal zone along the main lanes of fully interchanged highway or along the acceleration lane merging to the highway from the ramp shorter object values should be considered (0.15 m or 0.0 m).

Typical scenarios for implementation of DSD:

• Along freeways’ alignment.

• On fully interchanged highways, while approaching the exit ramp.

• Along the ramp in order to notice the traffic regulations and the interchange configuration prior to merging to the major highway.

• Along the ramp, approaching a signalized intersection if there is a change in lane configuration in the intersection leg.

SSD should be implemented along ramps prior to an intersection without increasing the number of lanes in the intersection approach.

Sight distance restrictions by vertical and horizontal curves along the ramp alignment, and also by bridge structures along the interchange zone should be considered.

SSD is the minimum sight distance that should be maintained along the ramp alignment. The SSD needs to be adjusted to longitudinal grade and ramp design speed.

Ramps that are connected to the minor intersecting highway of an interchange by an unsignalized intersection without additive lanes in the intersection approach shall be designed according to SSD. Since the driver is more alerted while driving along the ramp (compared to the interurban section) the SSD could be calculated in accordance with PRT (perception reaction time) of 2.0 seconds.

The object height for designing a crest vertical curve along the approach to an unsignalized intersection (without additive lanes) is 0.15 m.

The SSD model which is prevalent in many highway design guidelines (e.g. AASHTO [10], TAC [1], Austroads [3] ) is presented in Equation (2). The stopping sight distance model has two components: 1) the distance traveled during the driver’s reaction time; 2) the distance traveled during braking by implementing equivalent deceleration rate (d, meter/sec²).

The differences in SSD outcome are caused by its purpose and derived parameters.

SSD = t R 3.6 ⋅ V d + V d 2 2 ⋅ 3.6 2 ⋅ d (2)

where:

SSD—minimum stopping sight distance (m);

V_d—design speed (km/hr);

d—deceleration of passenger cars (m/s²) (Table 2);

t_R (or PRT)—perception reaction time (s), 2.0 seconds for ramps and ramp terminal applications.

Typical SSD design values (which depend on the ramp design speed) are presented in Table 2.

DSD is the basic sight distance for the design of intersecting highways at the

⁽¹⁾PRT of 2 section for SSD calculation on ramps. Design values of DSD are based on Israeli DSD model (Table 1, Bassan [7] [12] ).

interchange. Ramps that are connected to the minor intersecting highway of an interchange by a signalized intersection with additive lanes in the intersection approach shall be designed according to DSD. This situation is usually suitable for a ramp exiting from the major highway of the interchange and intersecting urban arterial (minor roadway of the interchange).

The object height for designing a crest vertical curve along the approach to a signalized intersection (with additive lanes) is 0.15 m. Typical DSD design values (which depend on the ramp design speed) are also included in Table 2, based on Israel model documented in Section 3.

The analysis of vertical curves on ramps towards the intersection of interchange minor road is accomplished for either crest vertical curve (ramp rises from the major highway) or sag vertical curve (ramp falls from the major highway), based on SSD (Table 3) or DSD (Table 4).

The radius of ramp vertical curve (crest or sag) according to driving comfort criterion depends on comfort centrifugal acceleration. The accepted centrifugal acceleration in vertical curves is 0.3 m/sec² in order to meet the comfort criterion.

The mechanical expression of this control is presented in Equation (3):

R v = V d 2 0.3 ⋅ 3.6 2 = 0.2572 ⋅ V d 2 (3)

R_v: minimum vertical radius based on driving comfort criterion (m);

V_d: design speed (km/hour).

The radius of ramp crest vertical curve according to safety criterion depends

Remarks: ^(1)The vertical radii for V_d = 30, 40 km/hour were derived by comfort criterion for radial vertical acceleration of 0.3 m/sec² ( R = 0.257 ⋅ V d 2 ) as presented by Equation (3). The values in parentheses are computed values derived by the SSD criterion (safety criterion). ⁽²⁾The vertical radii are suitable for passenger cars for SSD ≤ L and for tangents with opposite directions. For SSD > L and tangent with the same directions, relevant formulas (Equation (4), Equation (5)) shall be used [13] [14]. ⁽³⁾The two-lane, one-way ramp can be categorized as semi-directional ramp (V_d = 80 - 100 km/hour). ⁽⁴⁾Sag vertical curves were derived by comfort criterion (Equation (3)) assuming appropriate illumination.

Remarks: ^(1)The vertical radii are suitable for passenger cars for DSD ≤ L and for tangents with opposite directions. For DSD > L and tangent with the same directions, relevant formulas shall be used [13] [14]. ⁽²⁾The two-lane, one-way ramp can be categorized as semi-directional ramp (V_d = 80 - 100 km/hour). ⁽³⁾Sag vertical curves were derived by comfort criterion (Equation (3)) assuming appropriate illumination.

on the SSD (or DSD), driver eye height, object height, and the algebraic difference in grade (Equation (4) and Equation (5)). Equation (4) and Equation (5) are different if SSD (or DSD) is smaller or equal to the curve length (L), or if SSD (or DSD) is larger than the curve length (L):

R v = SSD 2 2 ⋅ ( h 1 + h 2 ) 2 for SSD ≤ L (4)

R v = 200 ⋅ SSD A − 20000 ⋅ ( h 1 + h 2 ) 2 A 2 for SSD > L (5)

R_v—radius of crest vertical curve (m);

A—algebraic difference in grade (percents);

SSD—stopping sight distance for design (m). DSD can be applied too;

h₁—design value of driver eye height (1.05 m for passenger cars);

h₂—design value of the object height: 0.15 meters for ramps towards intersection (for SSD, DSD) or along the major highway towards exit ramp (for DSD); 0.0 meters on entry ramp terminal along acceleration lane (for DSD).

Collisions on ramps and connecting roads generally increase with traffic volume and with decreasing curve radius [1] [4]. Additionally, since upgrade exit ramps have lower collision rates, it is preferable from safety point of view for the connecting road to pass over the freeway or major highway (higher speed road) as presented in the upper part of Figure 2.

If a significant decrease in traffic volume is expected along the major highway after passing an exit ramp, such that two basic lanes (instead of 3 basic lanes) could accommodate the traffic volume (for at least 2 kilometers), without a deterioration in LOS (C for a freeway in level terrain, otherwise D), then it might be possible to reduce one basic lane. The intention is that after merging in the same

⁽¹⁾The vertical radii are suitable for passenger cars for DSD ≤L and for tangents with opposite directions. For DSD > L and tangent with the same directions, relevant formulas (Equation (4), Equation (5)) shall be used [13] [14].

interchange, the difference between traffic volumes (exit and entrance) will still sustain as a traffic volume of one lane.

In principle the reduction of a basic traffic lane shall be applied after the physical nose as follows:

Decision sight distance (DSD) should be maintained between the edge of the painted nose (point C) until the beginning of the taper which directs the basic lane reduction. Figure 7 introduces a DSD of 325 m for a design speed of 110 km/hour in the major highway.

A traffic sign for lane reduction (enumerated as 621 in Israel) will be posted 150 meters prior to the beginning of basic lane reduction. For DSD = 325 m (derived by V_d = 110 km/hour), the traffic sign is located 85 meters after the physical nose.

The number of arrows shown in traffic sign number 621, will be identical to the number of lanes downstream, i.e. the number of legs will be equal to the number of lanes prior to lane reduction.

The taper for lane reduction will have an offset of 1:100 (length of 360 meters). The “reduced” lane is the rightmost lane.

A reduction of a basic lane prior to a two-lane exit ramp terminal is not recommended on freeways or interchanged highway. The logic is a bottleneck potential along the influence zone of basic lane reduction (approximated length of 360 m). This influence zone of basic lane reduction should be suitable for completing successfully the weaving maneuver which has two simultaneous traffic actions: 1) passing from the right lane to one of the left lanes for the vehicles driving along the major highway; 2) passing from the left lanes to the reduced basic lane (practically the exit lane to the right) for drivers exiting through the exit ramp terminal. DSD in such a scenario will be managed by directional and message signs which are posted prior to the beginning of basic lane reduction (Figure 8). This geometric design solution could be implemented in constrained situations, practically intermediate stage of construction or right of way limitations which are typical in Israel.