An unmanned system (UAV) has been designed for an autolanding capability. The UAV should be able to navigate and fly autonomously from any point within a defined three-dimensional (3D) space to a landstrip and land there without human intervention.
System testing provides answers to various questions about how well the system meets the specified requirements. These questions include:
1. Does the system design meet specified system performance?
2. If the system is produced in quantity, what is the percentage of produced systems that fail to perform as specified?
3. Under what conditions will the system continue to perform its function, even when used outside of specified environmental parameters?
4. Will the system meet its specified performance throughout its lifetime?
System requirements usually specify a range of environment conditions over which the system must operate according to the performance measures listed. Also, the system must meet its performance requirements over the lifetime (when system is brand new, as well as when it has aged).
System performance testing must give appropriate attention to all the above requirements if the tests are to be realistic and unbiased for the deployed system. The tests shall be planned according to the following considerations:
• Several different UAV replicas shall be tested.
• The environment conditions (temperature, wind velocity, precipitation, etc.) shall be varied over the specified ranges.
• Maintenance shall be performed in accordance with specified procedures.
• Selection of which UAV will be used for which test condition shall be entirely random.
The UAV autolanding design capability must be tested by bringing the UAV to any location within a 3D space in front of the landing strip, and by initiating the automatic landing sequence. System testing considers three input factors (X, Y, and Z) representing the initial location of the UAV in space and one output which indicates a Test Success Score (TSS). TSS is a continuous variable representing either total success (TSS = 1), partial success (0 < TSS < 1), or complete failure (TSS = 0). TSS is computed based on the UAV touchdown rate of descent, UAV angles (pitch, roll, yaw) relative to landing strip centerline and speed as well as landing strip locations of touchdown as well as completion of UAV rolling. A failed test is declared if (a) the ground operator has to abort the UAV autolanding sequence and manually control it or (b) the UAV either touches down or completes its rolling run outside the emergency areas of the landing strip, or (c) the UAV has been damaged in the landing process.
The number of possible tests for this problem is, for all purposes, very large, and the cost of each test is considerable. The tests are limited by defining a specific set of values for each factor, or defining a set of rules for determining these values. Total number of alternatives for each factor are given in the table.
UAV initial location (feet)
Factor Minimum Maximum Step size Number of alternatives
UAV – X 30 50 10 3
UAV – Y -20 20 10 5
UAV – Z 5 35 10 4
Deliverables:
1) Draw charts to represent the position of the UAV in space.
2) Determine the number of tests need to be performed according to the number of alternatives.
3) Generate two TSS values for each test based on the position UAV initial location.
4) Consider only the minimum and maximum location for the UAV and build a design of experiments analysis to identify the significant factors.
5) Perform an analysis of variance (including the error estimation, sum of squares, and significance