Sage Journals: Discover world-class research

Abstract

Different from traditional ground vehicles, planetary robotic rovers with limited weight and power need to travel in unfamiliar and extremely arduous environments. In this paper, a newly developed four-wheel-rhombus-arranged (FWRA) mobility system is presented as a lunar robotic rover with high mobility and a low-weight structure. The mobility system integrates independent active suspensions with a passive rotary link structure. The active suspension with swing arms improves the rover's capacity to escape from a trapped environment whereas the passive rotary link structure guarantees continuous contact between the four wheels and the terrain. The four-wheel-three-axis rhombus configuration of the mobility system gives a high degree of lightweight structure because it has a simple mechanism with the minimum number of wheels among wheeled rovers with three-axis off-road mobility. The performance evaluation of the lightweight nature of the structure, manoeuvrability and the mobility required in a planetary exploring environment are illustrated by theoretical analysis and partly shown by experiments on the developed rover prototype.

Keywords

Lunar Robotic Rover Mobility Manoeuvrability Lightweight Rhombus-Arranged Wheels

1. Introduction

Returning to the Moon for science exploration and building permanent bases on the Moon are vitally important steps in the development of outer space resources and the expansion of living space for human beings on earth [1 –3]. In early 2004, NASA announced its new plan to return astronauts to the Moon in 2018. Some nations [4 –6] are working on their own scientific missions to the Moon. European, Japanese, Russian and Indian space agencies have opened their new development plans in lunar exploration for the next decade. Their near-term goals focus on several missions to explore the Moon by using unmanned robotic rovers [7 –9]. China has begun the Chang'E programme for lunar exploration [6], which is a three-stage plan to launch a lunar orbiter, deliver a soft-landing lander with an unmanned rover, and take sample and return. With the success of launching the lunar orbiter (Chang'E-1 satellite) in October 2007, China's lunar objectives have attracted the world's attention. The second stage of exploration using an unmanned robotic rover to survey the lunar surface is scheduled in 2013.

The surface of the Moon is covered with fine and dry soil called regolith and the gravity acceleration on the Moon is about one sixth of that on the Earth [10, 11]. Mission failure sometimes occurs, for example the Russian Lunokhod 2 encountered difficulty when its wheel sank into the soft lunar terrain near a crater with a depth of 200mm. In order to carry out the mission to survey a large area of the lunar surface, the unmanned robotic rover is required to safely travel through unfamiliar terrain over a long distance and accomplish its exploration task under some environmental conditions such as dusty, soft terrain with rough rocks, a low-gravity environment and extreme differences in temperature [12 –15]. It should be mentioned that, unlike the Apollo Lunar Roving Vehicle driven by astronauts for Apollo missions 15, 16 and 17 [16], high mobility with multi-axis properties are usually essential requirements for an unmanned lunar rover. In addition, rovers are also weight and power limited, subject to space limitations and the launching cost of the rockets. Therefore, the following basic requirements of an unmanned robotic lunar rover make its design challenging:

(1) Simple mechanism with high reliability

(2) Lightweight for low launching cost

(3) High degree of mobility in very rough terrain

(4) Low power consumption for long-term exploration

(5) Small size for the limit of loading space of rocket

To meet the challenge, some researchers have been devoted to investigating the terramechanics mechanism of wheel-soil contact problems [12 –14] or mobility [15]. Grand et al. [17] gave a general formulation of the kinetostatic model of articulated wheeled-legged rovers and investigated the high stability and traction performance during motion on rough terrains. Mobility in very rough terrain is often very limited due to the absence of adequate locomotion concepts [18]. Many scientists have earnestly studied and developed new suspensions or new mobility systems [7, 18 –23] to realize a rover that is able to move on an unknown lunar or planetary surface. In particular, some new concept designs are needed to break through the technology bottleneck of high mobility and the lightweight nature required of a lunar rover. For example, Kubota et al. [8] proposed a new mobility system with five wheels, which is a simple and light mechanism like a four-wheeled rover and provides a high degree of mobility like a six-wheeled rover. The surface mobility of the ExoMars rover developed by ESA [24] is achieved through three pairs of wheels. Each wheel pair is suspended on an independently pivoted bogie and each wheel can be independently steered and driven. All wheels can be individually pivoted to adjust the rover height and angle with respect to the local surface in order to improve the environment. So far, the top-level rover among the ones launched successfully is NASA's Curiosity [25]. This is a car-sized robotic rover that is 2.9m long by 2.7m wide by 2.2m in height and has a mass of 899kg including 80kg of scientific instruments. Dozens of China's lunar rover prototypes have been developed by nearly 30 organizations. For example, the China Aerospace Industry Corporation developed a six-wheeled rover with Rocker-bogie suspension [19]. Harbin Institute of Technology in China investigated a series of rovers with eight wheels, six wheels and even two wheels. However, subject to the limitation of the launching cost and loading capability of the rockets, a trade-off between a lightweight structure vs. high mobility is preferable for the ChinaChang'E programme of lunar exploration. In general, a planetary robotic rover is designed subject to the requirements of a certain practical engineering project.

This paper reports on a new low-weight and high-mobility system with a four-wheel-three-axis rhombus configuration and the integration of independent active suspension and a passive rotary link structure. The rover equipped with the new mobility system is called the FWRA rover for short. This paper is organized as follows. A system overview of the newly developed FWRA rover is first presented. Then a theoretical analysis of the FWRA mobility system is presented, including its turning and climbing capabilities as well as lightweight structure. Finally, some experiments are used to validate the mobility performance of the developed rover prototype.

2. Four-wheel-rhombus-arranged lunar rover

Our objective is to provide a lightweight and high-mobility rover to be a candidate for China's future lunar exploration. The developed FWRA rover and its four-wheel-rhombus-arranged mobility system are illustrated in Figure 1. In order to maximize mobility performance and minimize the complexity of the system, the main characteristic of the FWRA rover is the new design concept in which the “four-wheel-three-axis” mobility system combines the active suspension with the passive rotary link structure.

The rover has a length of 1300mm, a width of 1100mm and a height of 600mm. The mobility system has a mass of only 17kg and a capability of bearing a payload of 120kg in the gravity field of the Earth. Four independent drive wheels are equipped to improve the off-road performance. Those wheels are controlled by an EC motor. The average moving speed of the lunar rover is about 1.7cm/s and the speed is adjustable from 0cm/s to 5.6cm/s. The developed rover is turned by both a front-and rear-wheel steering mechanism. Owing to such a simple steering mechanism, the rover is easily controlled to implement the turning function.

Figure 1.

Overview of the FWRA lunar rover: (a) physical prototype; (b) mobility system; (c) mechanism principle

Notice that very few types of materials are available for the purpose of lightweight structures for planetary rovers. To achieve a very lightweight structure, we have to resort to new concept designs of structures. In comparison with the existing six-wheeled mobility system, the FWRA mobility system in Figure 1 has not only the same three-axis off-road mobility as the six-wheeled ones but also the predominance of lightweight structures by reducing two sets of wheeled systems.

Due to the notable redundancy of structure mass, the FWRA mobility system can give a greater design abundance of optimization designs to upgrade its performance. For instance, the wheels with enveloped hubcaps in Figure 2(a) are used to avoid failure such as when one wheel of the well-known Opportunity Mars rover [21] was blocked by an inserted potato-sized rock due to the unsealed hubcaps, as shown in Figure 2(b). In our design, the material of the wheel tread and hubcaps is titanium alloy but the material used in the thin plates to cover the two sides of hubcaps is solimide polyimide foam insulation or phenolic impregnated carbon ablator.

3. Performance analysis of hybrid of active and passive suspensions

As shown in Figure 1, the FWRA mobility system consists of a front carriage and a rear carriage with four swing arms. The swing arms are connected with four wheels, respectively. The front carriage and rear carriage are coupled by a passive rotary pivot with an electric magnetic clutch. The front carriage and rear carriage are independent while the electric magnetic clutch is powered off and the mobility system presents the characteristics of passive suspension. On the contrary, the front carriage and rear carriage are integrated as a whole while the electric magnetic clutch is powered on and the mobility system shows the characteristics of active suspension. In most cases, the FWRA mobility system usually shows passive characteristics, which leads to continuous contact between the wheel and the rough terrain. The active-suspension characteristic is used only while the lunar rover encounters obstacles by accident or needs to escape from a trapped environment.

Figure 2.

(a) wheel of the developed FWRA rover; (b) failure case of Opportunity Mars rover.

In order to improve the escape ability of the rover, the angle of the swing-arm is controlled by a motor equipped with an anti-brake. Generally, the anti-brake is power off and the angle of swing-arm is fixed. Under a certain condition such as climbing over an obstacle or travelling in a crater, the anti-brake is power on and the swing arm can be used to achieve the required position by the motor. Although the motors in swing arms need more power than the wheel-drive mechanism do, no extra strong-power supply system is required. This is because: (1) the motors in the swing arm do not need to frequently keep working in general and the function of swing arm is used only for escape and (2) while the motor in a swing arm starts to work in order to increase the rover's escaping ability, only one drive motor of the wheel connected with the working swing arm needs to accompany the work. The other eight motors do not work at the same time.

3.1 Capacity of climbing-obstacle

Due to the swing arms, different strategies can be adopted for the rover to climb the obstacles. As usual, there are three steps when the rover climbs a vertical obstacle. In order to understand the climbing capability of the mobility system, the parametric models for climbing two types of obstacles are proposed, as shown in Figure 3. The sets of equations for the different steps are put forward based on the quasi-static analysis method. By using these equations, the minimum road adhesive coefficient at different heights (h), which is required by the wheeled model as the rover climbs the obstacle, can be obtained using this method. Then, by comparing the road adhesive coefficient, the climbing capability of the whole rover can be evaluated in theory.

Figure 3.

Parametric models for climbing obstacles remarked as Type-A and Type-B. For these angles, α,β,γ, the positive direction is anticlockwise

Step1:

{\begin{cases} G l_{04} + M_{3} + F_{3} h_{03} - (N_{3} - G_{w}) l_{03} = 0 \\ (N_{1} \sin θ + F_{1} \cos θ - G_{w}) l_{10} - (N_{1} \cos θ - F_{1} \sin θ) h_{10} \\ + 2 F_{2} h_{02} - 2 (N_{2} - G_{w}) l_{02} + M_{1} + 2 M_{2} = 0 \\ N_{1} \cos θ - F_{1} \sin θ - F_{3} - 2 F_{2} = 0 \\ N_{1} \sin θ + F_{1} \cos θ + 2 N_{2} + N_{3} - 4 G_{w} - G = 0 \end{cases}

(1)

Step2:

{\begin{cases} G l_{04} + M_{3} + F_{3} h_{03} - (N_{3} - G_{w}) l_{03} = 0 \\ (N_{1} - G_{w}) l_{10} + F_{1} h_{10} - 2 (N_{2} \cos θ - F_{2} \sin θ) h_{02} \\ - 2 (N_{2} \sin θ + F_{2} \cos θ - G_{w}) l_{02} + M_{1} + 2 M_{2} = 0 \\ F_{1} - 2 (N_{2} \cos θ - F_{2} \sin θ) + F_{3} = 0 \\ N_{1} + 2 (N_{2} \sin θ + F_{2} \cos θ) + N_{3} - 4 G_{w} - G = 0 \end{cases}

(2)

Step3:

{\begin{cases} G l_{04} + M_{3} - (N_{3} \cos θ - f d p_{3} \sin θ) h_{03} - (N_{3} \sin θ \\ + F_{3} \cos θ - G_{w}) l_{03} = 0 \\ (N_{1} - G_{w}) l_{10} + F_{1} h_{10} - 2 (N_{2} \cos θ - F_{2} \sin θ) h_{02} \\ - 2 (N_{2} \sin θ + F_{2} \cos θ - G_{w}) l_{02} + M_{1} + 2 M_{2} = 0 \\ F_{1} + 2 F_{2} - (N_{3} \cos θ - F_{3} \sin θ) = 0 \\ N_{1} + 2 N_{2} + (N_{3} \sin θ + F_{3} \cos θ) - 4 G_{w} - G = 0 \end{cases}

(3)

In Eqs. (1)–(3), F_i = µN_i is the tractive force of the ith wheel, N_i is the normal force of the ith wheel, M_i is the torque of the ith wheel with the expression M_i = µN_ir, G_w is the weight of the wheel, G is the weight of the payload, µ is the minimum road adhesive coefficient required by the rover while climbing the obstacle, r is the radius of the wheel, θ is the slop angle of the obstacle, α is the front swing-arms angle, β is the middle swing-arms angle, γ is the rear swing-arm angle, h_pq with h_pq = y_q – y_p, p, q, q = 0,1,2,3,4, is the vertical distance from the key point p to the key point q and l_pq with l_pq = x_q – x_p is the horizontal distance from the key point p to the key point q.

In general, the initial swing angles of swing arms are all zero such that each wheel evenly bears the weight of the payload. However, the angles of swing-arms (α,β,γ) are adjustable, which can influence the capacity of climbing obstacles. Therefore, it is necessary to make a qualitative analysis of the influence of the swing angles on the climbing capability. Strategic decisions can be made for the rover to climb an obstacle like type-A or type-B accordingly. To study the climbing capacity of the rover under severe conditions, the height of the obstacle is set to be more than the radius of the wheel and the angle of the obstacle (θ) is set to be zero (i.e., vertical obstacle).

Figure 4 shows that the relationship between the road adhesive coefficient and the height of the wheel (h) are varied in different steps. In Step 1 and 2, the road adhesive coefficients decrease with an increasing height of the wheel. The friction coefficients increase as the height increases in Step 3. The road adhesive coefficient must meet the requirements of all steps when climbing an obstacle. Thus, the top friction coefficient at different heights of the wheels should be chosen as the criterion to judge the capability of climbing an obstacle and it can expressed as μ_p = max(μ_m(h)). When the value of µ_p is bigger, the climbing capability of the rover is more difficult.

Figure 4.

Road adhesive coefficient (µ_m) vs. height (h) of the wheel in different steps: (a) Type-A, α = 0,β = 0, γ = 0, (b) Type-B, α = 0,β = 0,γ = 0

According to the analysis above, it can be seen that it is most difficult to climb the obstacles like type-A and type-B in Step 2. The comparison of the maximum friction coefficient indicates that it is more difficult for the rover to climb an obstacle like type-B than an obstacle like type-A.

Figure 5.

Influence of swing angle of arms on the peak value of road adhesive coefficient µ_p in different steps while climbing obstacle like type-A: (a) β = 0, γ = 0; (b) α = 0, γ = 0; (c) α = 0,β = 0.

Figure 6.

Influence of swing angle of arms on the top value of road adhesive coefficient µ_p in different steps while climbing obstacle like type-B: (a) β = 0, γ = 0; (b) α = 0,γ = 0; (c) α = 0,β = 0.

The influence of the front swing-arm angle α, the middle swing-arm angle β and the rear swing-arm angle γ on the capability of climbing an obstacle are investigated. Figure 5 and Figure 6 show the maximum values of the road adhesive coefficient (µ_p) at different angles of swing arms (from −30° to 30°), while the rover climbs the type-A obstacle and the type-B obstacle, respectively. From Figure 5(a) and Figure 6(a) it is evident that the angle of the front swing-arm (α) has a great influence on the peak value of the road adhesive coefficient µ_p in Step 1 but a smaller effect in Step 2 and Step 3. The values of µ_p increase as the angle of front swing-arm α increases. It can be seen from Figure 5(b) and Figure 6(b) that the angle of the middle swing-arm (β) has a remarkable influence on the peak value of the road adhesive coefficient of different heights (h) in Step 1 and Step 2 but only very little effect on the value of µ_p in Step 3. The value of µ_p in Step 1 rises with the increment of β. Instead, the value of µ_p in Step 2 decreases with the increment of β. Figure 5(c) and Figure 6(c) show that the values of µ_p in Step 1 and Step 2 decrease with the increment of γ. In addition, the value of µ_p in Step 3 decreases with the increment of the value of γ.

According to the analysis above, the minimum value of µ_p does not appear in the original state (α = 0,β = 0, γ = 0). Thus, it is very useful for the improvement of the rover's capability to climb an obstacle to investigate the influence of swing angles on the highest value of the road adhesive coefficient of different heights (µ_p). Due to the independence of the swing arms from each other, we can make different strategic decisions aimed at different obstacles. Figure 5 and Figure 6 indicate that the minimum value of µ_p can be obtained by individually adjusting the swing-arms. Therefore, through the combined adjustment of the swing arms, a better performance of the rover can be achieved.

3.2 Analysis of steering performance

As shown in Figure 1(b), the FWRA mobility system has only two steering mechanisms, which are attached to the front carriage and the rear carriage, respectively. It is easy to control the front steering mechanism in agreement with the rear steering mechanism. The front and rear wheels can independently turn around their joints from 0 to 360 degrees.

To meet with the requirement that all tires should roll without lateral sliding, the wheels should follow the curved paths with a different radius originating from a common turn centre [26].

Let B stand for the middle-wheel tread and R' be the turning radius of the rover. The FWRA mobility system can achieve three types of turning modes including big-radius turning (R' > B/2), small-radius turning (R' < B/2) and even zero-radius turning (R' = 0). The corresponding geometrical principles of these turning motions rover are shown in Figure 7.

Figure 7.

Parametric models for turning motion: (a) $R^{'} < \frac{B}{2}$ ; (b) $R^{'} > \frac{B}{2}$ ; (c) $R^{'} = 0$ (zero-radius turning).

Suppose that the steer angle in an anti-clockwise direction is positive. From Figure 7, the proper relationship between the steer angle of the front wheel α'₁ and that of the rear wheel α'₃ can be established as follows.

\cot α_{1}^{'} - \cot α_{3}^{'} = \frac{R^{'} - r_{s}}{L / 2} + \frac{R^{'} + r_{s}}{L / 2} = \frac{4 R^{'}}{L}

(4)

where r_s is the distance between the steering joint axle and the centre of the wheel and L is the wheelbase.

The linear speed V'_j of the jth wheel should satisfy the following relationship:

\frac{V_{j}^{'}}{r_{j}^{'}} = constant

(5)

where r'_j is the turning radius of the wheel j around the instant centre o of the velocity of the whole rover and j=1,2,3,…,6.

If a rover is trapped in a limited space surrounded by rocks or a crater, it must have the capability of escaping from there. The front and rear wheels can independently turn around their joints from 0 to 360 degrees, such that the FWRA mobility system has an escaping ability of zero-radius turning, as shown in Figure 7(c). The property of zero-radius turning implies that the rover can escape from any direction of the trapped environment.

From the analysis above, it can be seen that only synchronization between the two wheels can complete the turning of the whole rover. Due to the simple mechanisms, the mobility has the characteristic of high reliability and is easy to operate. This characteristic is one of advantages for the FWRA mobility system in comparison with the four-wheel turning synchronization of the well-known Rocker-bogie suspension [19] or the synchronization between the steering of the front and rear wheels and the speed difference of the bogie wheels in Shrimp design [18].

3.3 Function of active suspension

Active suspension with swing arms is a key component to improve the stability and escaping ability of the lunar rover. One of the properties of the FWRA rover is the adjustability of the centre of gravity. There are two ways to improve the anti-roll-body performance using the swing arms when the rover travels on the slope. One is to decrease the centre of gravity and the other is to adjust the body's heeling angles, as shown in Figure 8. Thus, the lateral stability of the rover can be improved by adjusting the swing arms if necessary. There is a trade-off of ground clearance vs. stability in practical engineering.

The adjustability of the centre of gravity by using active suspension with swing arms can bring a new type of escaping ability for the FWRA rover. Assume that the rover is designed so that each swing arm has a proper slope such that the rover, in a normal state, can travel with the design-required ground clearance but not with the maximum one. If the moving process of the unmanned rover is blocked by a rock with a size close to the current ground clearance, the slope swing arms become upright by rotation and the ground clearance is then raised such that the blocked rover can back off from the obstacle. Contrastingly, for a rover equipped only by completely passive suspension, it is very difficult or even impossible to escape when the rover is blocked by a ground-clearance-size rock by accident.

Figure 8.

Improvement of stability: (a) normal condition, (b) adjusting the swing arms to lower the centre of gravity h<H, (c) adjusting the swing arms to keep the body's balance

Another remarkable property of the FWRA rover is that with the operations of turning the front or rear wheel, or swinging the arms of the middle wheels, the projection of the gravity centre of the rover may lay in the triangle surrounded by any three wheels as we desired. Notice that the contact-ground points of wheels are changed but the centre of gravity is not transferred.

Figure 9.

The particular stability property of the FWRA mobility system

The property shown in Figure 9 brings about a three-axis and three-track escaping ability. For instance, when the front wheel sinks into a soft-soil hole or a V-shape slot between two hard stones, the electric magnetic clutch in Figure 1(b) firstly powers on such that the front carriage is rigidly connected with the rear carriage. Next, by using the swing arms of the middle wheels, the projection of the centre of mass can be placed in the rear triangle. Finally, the front wheel can be freely turned up by the swing arm to escape the trapped environment. In addition, by using the independent turning-mechanism, the projection of the centre of mass may lay in the left triangle or the right triangle. These functions can help the lunar rover escape from various severe conditions such as a high step with a height of 280mm and cross a rift with a width of 350mm more than the diameter of the wheel. The details are demonstrated in the physical experiments in Section 4

It should be stressed that for the FWRA mobility system, the active suspensions are used only as the rover encounters obstacles by accident or needs to escape from a trapped environment. Usually, this rover moves simply by way of passive suspensions.

3.4 Performance comparison with other similar system

It should be mentioned that there is an obvious disadvantage of stability for the FWRA mobility system. That is, the projected contact ground area of its four rhombus-arranged wheels is smaller than that of the four or six rectangular-arranged wheels of the same size. This property brings the problem that for lateral stability, the minimum rolling moment arm OB of the FWRA mobility system is slightly smaller than that of the wheel-rectangular-arranged mobility system marked by OA, as shown in Figure 10. Fortunately, different from the traditional vehicle dynamics of a car, the speed of a lunar rover running on the lunar surface is very slow. For example, a mean speed of 5cm/s is acceptable for an unmanned lunar rover in China.

Figure 10.

Stability comparison between the FWRA mobility system and the ones with four or six rectangular-arranged wheels

With respect to mobility and escaping ability, a mobility system with four rectangular-arranged wheels, even equipped with active suspension, such as the redundant wheel-legged robot Hylos [22] or the 4sRR OpenWHEEL robot [23], is obviously without the three-axis-and-three-track escaping ability of the FWRA mobility system. Moreover, it is very difficult for the rectangular-arranged wheel mobility systems to climb a step with a height that is more than the radius of the wheel and it is also impractical for crossing a rift of the width more than the wheels' diameter. For example, the well-known Rocker-bogie mobility system with six wheels could not cross a wheel-diameter width rift.

The proposed four-wheel-rhombus-arranged mobility system is similar to the six-wheel Shrimp rover [18], which also uses the design of a rhombus configuration. For the six-wheel Shrimp rover, more wheels provide better stability. However, in order to minimize slippage while turning, it requires a remarkably complex control during its turning manipulation because of its turning mechanism, which is necessary to synchronize the steering of the front and rear wheels and the speed difference of the bogie wheels. In addition, it is obvious in terms of the structure's weight that the proposed mobility system is a better choice than the Shrimp rover.

4. Experimental Test of Physical Prototype

A well functioning prototype of the FWRA mobility system has been designed and manufactured. A lot of experimental tests including some general cases and escaping cases have been carried out on the lunar simulation terrain to validate the predicted performance. The image sequences obtained from the experiments are shown in Figures 11–14. During the experiments, the developed prototype encountered and traversed a big crater and a rift with a width of more than a wheel diameter, scrambled over rocks and a 280mm high step, climbed a soft-sand slope up to 30 degrees, withstood a tilt of 40 degrees in any direction without over-turning, avoided obstacles by turning with zero radius and so on. The experimental results show a good mobility performance.

Figure 11.

Experiments in general cases: (a) anti-rolling in a slope, (b) climbing a slope up to 30 degrees, (c) scrambling over rocks, (d) zero-radius turning, (e) travelling through the crater, (f) adjusting posture of the rover body to improve the antenna direction to the Earth.

Figure 12.

Various stages in climbing a step with the height of 280mm

Figure 13.

Various stages in multi-step-crossing a rift with the width of 380mm and the rift width is more than the wheel's diameter of 330mm

As an example, the detailed experiment process in crossing a rift with a width of 380mm in Figure 13 is described as follows: as the FWRA rover detects or approaches a rift, the CCD camera can capture a picture of the rift and the image recognition technology of the in-vehicle computer can compute the width of the rift. If the width of the rift is smaller than 300mm, a decision on implementing the direct-crossing procedures can be made. Only if the rift width is more than 300mm a decision on implementing the multi-step-crossing procedures should be made based on the environment information obtained by the CCD camera. The corresponding path planning is figured out for the GNC (Guidance, navigation and control) system. Notice that the GNC system of the FWRALR is related to the dual modes including visual-based navigation technology and remote control (earth-moon communications) by teleportation. Using the GNC system, the electric magnetic clutch firstly powers on such that the front carriage is rigidly connected with the rear carriage. Then, the arms of two middle wheels may swing ahead with a proper angle such that the centre of gravity of the rover is located in the triangle area surrounded by the rear wheel and the two middle wheels (see, Figure 9(a)). In what follows, the front wheel can directly move through the rift. As the front wheel passes through the rift, the two middle wheels approach the bank of the rift. The operations of right-direction turning of the rear wheel is used to locate the centre of gravity in the triangle area surrounded by the front, rear and left-side wheels, as shown in Figure 9(b). The right-side wheel easily crosses the rift. Similarly, the left-side wheel can cross the rift by turning back the rear wheel and turning the front wheel to the left-side direction. After the two middle wheels pass the rift, the swing arms of the middle wheels can swing back in the rear direction such that the centre of gravity is located in the triangle area surrounded by the front, two middle wheels (see, Figure 9(d)). Finally, the rear wheel can easily pass the rift without any difficulty and the process of crossing the rift is finished.

The other experiments included vibration mode testing, analyses of potential failure cases and so on. Figure 14 shows a potential failure case and its solution. Suppose that the drive motor of a certain lateral wheel fails. We firstly make the failed wheel be a rotatable idler. If there is no yaw control for this failure case, the rover has a drift of 2.6m as the path deviation when it travels a straight-line distance of 10m, as in Figure 14(a). There are two solutions of yaw control for the rover to solve this potential failure case. One is to give up the drive function of the other lateral wheel to make it also become a rotatable idler. In our design, it is enough for the rover to only use the front and rear wheels as drive wheels. The other is to slightly turn the front wheel and keep three drive wheels. Figure 14(b) shows little drift when the rover with one failed lateral wheel travels a straight-line distance of 10m under the yaw controls as mentioned above. It should be mentioned that, when the drive motor of the front or rear wheel fails, there is no need to do anything if the failed wheel can run as an idler. This is because no path deviation will occur due to the symmetry of the whole mobility system and the sufficient drive power of three wheels. In contrast, it is necessary for a six-wheeled system like the Rocker-bogie mobility system to implement yaw controls to eliminate path deviation when one of the six drive wheels fails.

Figure 14.

Drive failure tests of one lateral wheel: (a) 2.6m drift within 10m distance of travel without control, (b) little drift with simple control

5. Conclusions and Remarks

A high-mobility and low-weight mobility system (called FWRA) for lunar robotic rovers is presented in this paper. The mobility system has only four wheels and is equipped with active suspensions with swing arms and integrates a passive rotary pivot. In a general way, the passive rotary link structure can guarantee continuous contact between the four wheels and the terrain without extra energy cost. Only as the rover encounters obstacles by accident or needs to escape from a trapped environment do the suspensions switch from a passive state to active function. Among wheeled rovers with three-axis off-road mobility, the FWRA mobility system has the highest degree of lightweight structure due to the minimum number of wheels. The four-wheel-three-axis rhombus configuration also gives a high escaping capability in unknown environments. The performance comparison with other similar mobility systems is also discussed.

Footnotes

10. Acknowledgments

The authors would like to thank the anonymous referees for their useful comments and suggestions,which lead to the improvement of this paper. This work was supported by the National Science Fund for Distinguished Young Scholars in China (No. 11225212),the National Science and Technology Support Programme (No. 2012BAH09B02),the Specialized Research Fund for the Doctoral Programme of Higher Education (No. 20120161130001),and the Hunan Provincial Natural Science Foundation for Creative Research Groups of China (No.12JJ7001).

References

Ehrenfreund

McKay

Rummel

J.D.

Foing

B.H.

Neal

C.R.

Masson-Zwaan

Ansdell

Peter

Zarnecki

Mackwell

Perino

M.A.

Billings

Linda

Mankins

Race

, (2012) Toward a global space exploration program: A stepping stone approach. Advances in Space Research. 49(1): 2–48.

Benaroya

Bernold

, (2008) Engineering of lunar bases. Acta Astronautica. 62(4–5): 277–299.

Siegfried

W.H.

(1999) Lunar base development missions. Acta Astronautica. 44(7–12): 755–767.

Matsumoto

Kamimori

Takizawa

Kato

Oda

Wakabayashi

Kawamoto

Okada

Iwata

Ohtake

, (2006) Japanese lunar exploration long-term plan, Acta Astronautica. 59(1–5): 68–76.

Sasaki

Kubota

Okada

Saiki

Kuroda

Kunii

Shibamura

Akiyama

Ohtake

Ichikawa

Higa

Hirata

Sugihara

Haruyama

Otake

Yoshioka

Terazono

Yamada

Yamaguchi

Kodama

, (2002) Scientific exploration of lunar surface using rover in Japanese future lunar mission. Adv. Space Res. 30(8): 1921–1926.

Zheng

Y.C.

Ouyang

Z.Y.

C.L.

Liu

J.Z.

Zou

Y.L.

(2008) China's Lunar exploration Program: Present and future. Planetary and Space Science. 56(7): 881–886.

Weisbin

C.R.

Lavery

Rodriguez

, (1997) Robotics technology for planetary missions into the 21st century. In Proceedings of the i-SAIRAS′97, pp. 5–10.

Kubota

Kuroda

Kunii

Nkatani

, (2003) Small, light-weight rover “Micro5” for lunar exploration. Acta Astronautica. 52(2–6): 447–453.

L.H.

Zhao

Y.B.

, (2011) Slip ratio based traction coordinating control of wheeled lunar rover with rocker bogie. Procedia Engineering. 15: 510–515.

10.

Khan-Mayberry

, (2008) The lunar environment: Determining the health effects of exposure to moon dusts. Acta Astronautica 63 (7–10): 1006–1014.

11.

Bui

H.H.

Kobayashi

Fukagawa

Wells

J.C.

(2009), Numerical and experimental studies of gravity effect on the mechanism of lunar excavations. Journal of Terramechanics. 46(3): 115–124.

12.

Thueer

Siegwart

, (2010) Mobility evaluation of wheeled all-terrain robots. Robotics and Autonomous Systems 58(5): 508–519.

13.

Nakashima

Fujii

Oida

Momozu

Kanamori

Aoki

Yokoyama

Shimizu

Miyasaka

Ohdoi

, (2010) Discrete element method analysis of single wheel performance for a small lunar rover on sloped terrain. Journal of Terramechanics. 47(5): 307–321.

14.

X.L.

Fang

Liu

J.F.

(2012) Interaction mechanical analysis between the lunar rover wheel-leg foot and lunar soil. Procedia Engineering. 29: 58–63.

15.

Schlutz

Messerschmid

, (2012) Integrating advanced mobility into lunar surface exploration. Acta Astronautica 75: 15–24.

16.

Morea

S. F.

(1988) The Lunar Roving Vehicle– Historical Perspective. In Proc. 2nd Conference on Lunar Bases and Space Activities, 5–7 April 1988, NASA Conference Publications 3166, 1: 619–632.

17.

Grand

Benamar

Plumet

, (2010) Motion kinematics analysis of wheeled–legged rover over 3D surface with posture adaptation. Mechanism and Machine Theory. 45(3): 477–495.

18.

Siegwart

Lamon

Estier

Lauria

Piguet

, (2002) Innovative design for wheeled locomotion in rough terrain. Robotics and Autonomous Systems. 40: 151–162.

19.

Chen

B.C.

Wang

R.B.

Jia

Guo

Yang

, (2009) Design of a high performance suspension for lunar rover based on evolution. Acta Astronautica. 64(9–10): 925–934.

20.

Flippo

Miller

, (2011) Suspension and wheel evaluation and experimentation testbed for planetary rovers. Int. J. Mechatronics and Automation. 1(1): 29–37.

21.

Leger

Trebi-Ollenu

Wright

Maxwell

Bonitz

Biesiadecki

Hartman

Cooper

Baumgartner

Maimone

, (2005) Mars exploration rover surface operations: Driving spirit at Gusev Crater. In: Proc. IEEE Conf. Syst. Man Cybern. The Big Island: Hawaii, pp. 1815–1822.

22.

Amar

F.B.

Grand

Besseron

Plumet

, (2004) Performance evaluation of locomotion modes of an hybrid wheel-legged robot for self-adaptation to ground conditions. Proceedings of the 8th ESA Workshop on Advanced Space Technologies for Robotics and Automation, Noordwijk, pp. 1–7.

23.

Fauroux

J.C.

Forlorou

Bouzgarrou

B.C.

Chapelle

, (2007) Design and Modeling of a Mobile Robot with an Optimal Obstacle-Climbing Mode. Proceedings of the 2nd World Congress in Design and Modelling of Mechanical Systems, CMSM′2007, Monastir, Tunisia, pp. 98.

24.

http://exploration.esa.int/science-e/www/object/index.cfm?fobjectid=45084.

25.

MSL at a glance - CNES. Cnes.fr. Retrieved 2012-08-07.

26.

Wong

J.Y.

(1989) Terramechanics and off-road vehicles, Elsevier: Amsterdam.