By Jerry Bradshaw, Training Director, Tarheel Canine Training, Inc.
Draft May 23, 2009
In the previous installment of this series, I discussed how to use the power of reward to improve the quality of the dog’s response during obedience. In the present article we will break down one of the more complex exercises in police K9 work, the building search, and discuss how reward can improve the quality of your building search work. As with any complex exercise, the key to training it properly is to deconstruct the exercise and break it down into its component parts, and train the parts. In doing so, we remove some of the variables that can impede the dog from thoroughly learning each of the components.
The building search can be broken down into three components: the start, the search, and the alert. If you are trying to train the alert, for example, and you make the search component of the training session too complex, you will make it more difficult for the dog to learn the alert. Removing unnecessary variables from a training session allows us to concentrate on the key concepts of the piece of the training we are attempting to teach the dog. Therefore, when teaching the alert, simplify the start and the searching components of the session.
Pre-requisite Exercise: The Alert on Command
The alert on command, which places the dog’s aggression on the command of the handler, is an exercise that is often ignored, but it may be your most important exercise in police patrol training. In much of the training I watch, the dog is brought to the training field, the decoy starts the action by agitating the dog, and the dog is sent to bite. The problem with this sequence of events is that the dog learns that his cue to get aggressive is the agitation rather than the command word of the handler to alert. The problem stays hidden, until the dog is deployed and the suspect has no bite equipment on, and is passive (sitting down or laying on the ground). In many cases the dog shows confusion and begins to look around at back up officers who may be standing up and moving around (his familiar context for bite work). The dog fails to engage, and rather than blaming the training as incomplete, the dog is blamed.
The proper sequence, which should be trained during each and every bite session from the very start of the dog’s training, is that the decoy always starts out passive, and the handler alerts the dog on the passive subject. Once the dog shows aggression, the decoy reacts to the aggression either by pressing the dog in defense, or by fleeing in prey, depending on the exercise. The aggression is placed on a variable reward system, where sometimes the passive decoy will flee after one bark, then the next time we would make the dog bark at the decoy for 15 or 20 seconds before motion or advance on the dog, and back to 5 or 10 seconds of barking, varying the amount of aggression required to bring the decoy alive. In the dog’s mind, he is bringing the passive person “alive” by his aggression. The decoy must reward this behavior at all times. Once it is done well, you can point your dog at anyone, give an alert command, and expect aggression and focus.
This teaches the dog aggression on command, rather than on the context (movement or threat) the dog perceives from the decoy. In fact, as the handler, you may perceive an apprehension situation well before your dog, and so if you can put the dog in an aggressive mood on command he will be ready to react immediately, and not be caught off guard. Work your bite sessions with the decoy in a hidden sleeve, sitting in a chair, or laying on the ground. You can work on a slick floor to slow the entry down, and set up furniture to protect the exposed areas of the decoy, and always send the dog on a long line to maintain positive control. The dog learns to alert on passive subjects in all positions, and becomes aggressive on your command.
A further extension of this exercise is to roll up to a strange place where you have already stashed a decoy behind the corner of a building or a dumpster, and take up a tactical position and give your alert command. Your dog may at first be confused, but be patient, and when he gives one bark, the decoy should jump out and flee, and in response send your dog for a reward grip. Do this in a bunch of different contexts with that one bark bringing out the decoy. Then ask for a few more barks, and then finally place this reward on a variable schedule so that the dog learns he has to sometimes bark for a while to get this result. The dog learns he can both turn a passive person alive by getting aggressive on command, and as well he can make people appear from behind objects and around corners when he becomes aggressive on command. I think you can see how this will help develop the start to your building search.
It is in this part of the behavior chain that I often see handlers struggling with success. One of the major complaints I hear is that the dog goes into the building, but actually starts searching for contraband rather than searching for a hidden suspect. Alternatively, the dog will go into the building but does not make the connection between entering the building and searching with determination, despite the fact that the dog seems to have a high hunting drive. The dog looks to the handler for help through the confusion and if you help him, he may come to rely on you going with him rather than searching independently.
In our program, we train the start of many searching exercises, including trailing, building searches, area searches, and detection, using the principle of HOT, WARM, and COLD. In this system, the initial cue that impels the dog to begin the search is systematically faded out until it is no longer present in the context of the exercise. Further, we isolate the component we are training, the start, from the other components, by simplifying the search and the alert portions of the exercise.
We begin by teaching the dog to search one room. We want the dog to have quick rewards for taking the start, so that the dog doesn’t get so wrapped up in the searching that the link between the start command/context and the reward is broken. Therefore the decoy is at first out in the open in various positions in the room, to the left and right of the entry door, in the corners, left and right of the entry door, sitting in a chair in the middle of the room, and then work up to being partially hidden but always accessible to the dog for a grip. Remember the grip is the reward for the start behavior, and we want to keep the search of the room itself very simple so that the dog concentrates on conditioning to the start rather than focusing on a complex search of the room. There will be plenty of time for that later.
The Hot Start
The dog is placed on a long line, and the handler takes a tactical position outside the door of the room we are going to search. The handler holds the dog by the collar. The decoy is already in the room, the handler gives an alert command to place the dog in an aggressive mood, and when the decoy hears the dog alert, he makes some agitation at the door to draw the dog to the room (note the sequence of alert first, and then agitation), and then the handler gives the command to apprehend, and releases the dog into the room, allowing the line to slide through his hands. Once the dog makes contact, proper line work is done, to set the grip and as well to encourage the dog to adjust full into the grip, and push the decoy. The handler keeps his tactical position, and then extracts the dog and decoy through the doorway, for a release, or tactical removal. The dog is removed and heeled away around the corner so the decoy can go back into the room, and the exercise is repeated, with the decoy being in a different place in the room. Do 4 or 5 repetitions, changing the decoy’s placement to encourage a good search pattern. I like to set the reward system of decoy placement up as follows: (1) Doorway, (2) along wall adjoining the doorway, (3) corner, (4) far corner, (5) opposite side of the doorway. Think of it as similar to making good hide placement in detection training. The decoy placement will draw the dog to search without blowing fast through doorways. By limiting the dog to one room initially, the dog will learn the habit of searching each room completely. As the training progresses to actual hiding places later, the dog learns to check these productive areas thoroughly.
Multiple repetitions are to be made. I suggest changing the angle of entry so the dog gets used to starting from either side of a doorway. If you have a room with multiple entry ways, that is great, you can start the dog from different angles into the room. Keep the search simple and do not ask yet for an alert, just work on the start with quick rewards. You can vary general variables such as floor surfaces and the degree of illumination in the room, for example.
The Warm Start
Once the dog seems to be clear on the hot start to make entry and complete a simple search of the room for a reward grip, it is time to move to what we call the warm start. The warm start is the mechanism by which we systematically fade out the cue of the agitation to impel the dog to make entry and search.
The dog is placed in the usual tactical position outside of the door. The alert is given, and some minimal agitation is given in response to the dog’s alert. But instead of releasing the dog into the room, the dog is praised for his alert, and quickly taken away from the set up position and walked away from the room out of sight of the doorway he will enter in a few seconds. Initially this break is only about 30 seconds, or just enough time to walk away and then come back. Once the dog is returned to the tactical position, the alert is re-given, but this time there is no further agitation from the decoy. Once the dog alerts the handler releases the dog into the room for a quick reward. For the first few warm searches, give the dog a quick reward bite right around the doorway so we make a rapid association from setting up and being released into the room without hearing the agitation.
A systematic lengthening of the time between the initial set up and the second set up and send into the room now takes place. We increase the time between the set-ups from 30 seconds to a minute, and then as we progress longer, we put the dog back in his vehicle while we wait for the time to elapse, working our way up to 15, 20, 30 minutes in between, up to an hour.
Look for a time threshold where the dog shows confusion. If you see confusion, say, after a 15 minute break, then lessen the time for a while and then work your way back toward 15 minutes. Then continue your progression. Typically this progression happens quickly. Initially, I will use the same room and entry door for a while as I progress through the warm starts. When I change the room being used, like when we use a different building, I will step back and reduce the wait time on the first search in a new area. Then you can rapidly increase the wait time back to the last threshold you completed successfully, and then beyond.
Now, remember, we have kept the search simple and asked for no alert. When the dog is making solid warm starts, you can begin to increase the complexity of the searches, using larger rooms, with still more (accessible) hiding areas. We still ask for no alert, we will train that separately, later. When the dog can do warm starts with wait times of 30 minutes to an hour, you can begin training the cold starts. This sounds like a lot of training time, but you will be surprised how quickly the dog makes progress, and further, because you are being systematic, there will be fewer times you will need to stop progress and remediate.
The Cold Start
The cold start is an extension of the warm starts. The decoy is placed immediately behind the door in the room. The dog is brought up to the tactical position outside the door. The dog is given the alert, and the handler releases the dog into the room. Notice there is no agitation at all. Once the dog enters the room, the reward bite is immediately there. I suggest you set up about 3 locations where a cold start can be made, with a quick reward around the corner from where the dog makes entry. A school is a great place to practice this part of the training. Once you see the dog making nice starts each time, you can then increase the search by moving the decoy deeper into the room, in a variable fashion. With the start progressed to this point we can now concentrate on increasing the complexity of the search.
The goal is to get the dog to search every room methodically. There are a number of variables that need to be addressed in the search portion of the training. These include the set time of the decoy, size of the search area, high and low searching, deep searching. In the section above on training the start, we kept the searches simple, with the decoy mostly standing or sitting at nose level, in a school classroom sized room, with the decoy placed very accessibly but not necessarily easily visible to the dog. Now it is time to vary these variables, but I caution you to vary them one at a time. If you change more than one variable, you work against success.
If you change the size of the room searched, keep the other variables the same. As the dog does a better job searching methodically in larger rooms, go back to a smaller room as you place the decoy high for the dog to locate, or prone or sitting on the floor. As you vary the depth of the decoy (behind furniture for example, or deep in an open closet door) reduce the size of the overall search area. At this point the dog is started cold every time.
In our academy we deploy the dog on a long line to restrict the search area to a room by room advance using a modified “clear, down and cover” approach. Without getting into a debate on the relative merits of long line versus off leash deployments, suffice it to say that you will get more thorough room by room search behavior, and a higher degree of certainty that a room was cleared by the dog thoroughly, if you condition the dog to search each segment by restricting his advance.
It is important to maintain the variable placement of the decoy during all this training. As the dog becomes comfortable with searching each room thoroughly, and making an advance to a new cover position for another deployment, we can start training the alert, but we do this separately from the search training.
Many times I see dogs search a building with determination, but they get to a hiding place where the subject is inaccessible, and the dog gives some change of behavior to indicate that he found the bad guy, but he doesn’t give a strong alert or even walks away from the hiding place. In this section I will concentrate on training an aggressive alert. However, there are many reasons to prefer a passive alert if you are doing a true clear, down, and cover building search. The aggressive alert is usually customary when dogs search buildings off leash, because if the dog gets deep into the building, you will have no way to know where he is if you don’t have some kind of audible alert. However if your dog is not allowed to range a great distance, and especially when doing a modified SWAT method of long line deployment, a passive alert preserves your hearing to detect any other movement (nobody commits a crime alone) and also preserves your tactical position. In many respects it is a much safer deployment.
Training the alert itself involves simplifying the search significantly, and you go back to hot starts during this phase of training to keep the dog’s drive in a high state when he encounters the decoy. I like to get a school building or industrial building with a long hallway with opposing doors. This process is very much like teaching a narcotics dog to alert on a hidden odor. Normally in this process of teaching an aggressive narcotics alert we do alert drills, which do not involve any complex searching, to increase the behavior of scratching when encountering an inaccessible odor hide. We do the same thing with the building search.
I start with the decoy half in one of the rooms close to the end of the hallway from which we are starting the dog. The handler once again gives his alert command to the dog, while the decoy is passive. On each bark, the helper moves closer to the dog, taking very quick discrete steps toward the dog, after each bark, up to the point where the dog is rewarded with a grip when the helper is within striking distance. I like to do this with a hidden sleeve. The decoy keeps the hidden sleeve behind his back, as he is drawn in by the dog. We want to keep this as realistic as possible. After the grip, the dog is worked properly, and disengaged. Keep the actual bites short so not to wear the dog out and reduce the barking from fatigue. The dog is then taken around the corner, and the decoy sets up across the hall in another room, half way in. The handler takes up position where the dog is looking down the hallway, and gives the alert command again, and the dog draws the decoy out of the room to him with each bark. This process is repeated 4 or 5 times, depending on the dog’s conditioning. Do not work the dog to the point of failure in the bark from fatigue. It is better to do short sessions with periods of rest in between, then drag out each individual session too long.
Once the dog is drawing the decoy out of each room in the hallway with a good continuous bark, the problem is then set up to begin the same way, and each time the dog barks the handler allows the dog to move discretely toward the decoy, by letting out a little bit of line after each bark. As the dog gets about half way to the decoy, the decoy moves just inside the door of the room he is starting in, leaving a very small open seam in the doorway. I post my foot against the bottom of the door and hold the handle with my hand pulling against my foot to steady it. I keep the bite arm as far from the door seam as possible, because I want the dog to sniff my pants leg that lies along the door seam, rather than alert to the smell of the bite equipment. As the decoy slips inside the door the handler allows the dog to run the rest of the way to the door seam. Now it is the decoy’s job to read the situation. Listen for sniffing behavior, and if you get a good change of behavior, reward that quickly just one or two times, by popping the door open immediately following a couple good sniffs, and giving the dog a bite. Make each grip quick, and then disengage the dog and move him out of the hall around the corner, so the decoy can set up again.
Set up the same way again, but the decoy goes across the hall, half way in the door. After a few good alert barks on the passive decoy, the decoy then slips behind the door, and sets up with a thin vertical seam in the door held between the handle and your foot at the bottom. The dog is given the apprehension command, and then allowed to go freely toward the door, handler loosely holding the long line, and this time, once the dog sniffs, wait for a bark or at least a squeak. Immediately reward any aggressive vocal behavior. Don’t try to get too much bark too soon, be patient. You can see the progression from here is to systematically increase the criteria required for the grip. We want to encourage the sequence: sniff, identify human odor, and then bark. Once we are getting a few good continuous barks, put the reward on a variable schedule.
Now we need to fade the visual cue of the decoy in the doorway. Again we use the sequence hot, warm and cold. If the visible passive decoy is considered a hot start, now we set the dog up to see the decoy run among a few of the doors, and then take the dog around the corner for a few seconds. While the dog goes around the corner, the decoy changes to his final position (use the first 4 doors for this shell game from the starting place of the dog) and hides in the fashion we have described, allowing a very thin seam in the door, with the decoy’s leg pressed against it.
After a while close the door all the way, keeping your pants leg along the seem, and your foot against the bottom seem of the door. The other rooms have some human odor in them. The decoy leaves the doors pulled almost closed, so the dog will want to check them, and possibly push the door open and investigate the room. Allow the dog to satisfy his curiosity. You have the 15 foot long line so you can help direct the search by casting the dog from door to door if he gets confused about what to do next. If the dog is searching hard (which he should be because we already developed this behavior so it should be strongly conditioned now) allow him to do it on his own. Keep quiet. Do not talk the dog into alerting on something. When the dog encounters the door with the decoy behind it, it is the decoy’s responsibility to reward the “sniff, identify, and bark” sequence. Because we now have made the search somewhat more complex, reduce the criteria for a reward. One bark gets a bite. Repeat this warm process as we did in the start procedure. This should be a very familiar process for the dog, and you should move through it quickly. Increase the criteria for the bark so you get a strong bark response from the dog.
Finally go to a cold start. On the cold start, again, reduce the criteria for a reward, and then rebuild it. It will rebuild quickly if you did your warm work thoroughly. Again a school is a great place to do this, because you can use 3 different hallways for the exercise each time you complete it. Some of the room doors during the cold starts should be closed and kept blank, so the dog doesn’t associate the find with a door he can’t push open. Treat blank rooms as you would blank boxes in narcotics alert training. If the dog tries to give a false response on a closed door to a blank room, give a reprimand, and move the dog away from the incorrect location and encourage him to resume searching. You can build the blank rooms to contain unused sleeves, and other unused equipment (containing no human odor) to proof the dog off of equipment as a reason to respond on the door, just as we proof the dog off of toys, food, baggies, etc. in narcotics training.
Build the Final Behavior Chain Complete
Now that we have built each component of the behavior sequence of the building search we can put it all together. Start small and make a quick reward when the dog solves the problem. We will do a cold start on a relatively small search problem (one room with a couple of doors for hiding places). The decoy should be hidden for at least 10 or 15 minutes prior to beginning. Pick a building with 2 or 3 entry ways from the outside, so we can do 2 or 3 repetitions with the decoy in the same hiding place, or else, change hiding places, but remember to open the door to the previous hiding place and know there will be a hot spot there you may have to help the dog overcome. He has encountered these hot spots before in the alert training so as long as he can investigate the area where the hot spot is, he will be able to source the decoy.
As you make the problems more complex, bear in mind a few things. As you vary the complexity of the alert, by hiding the decoy high, low, deep, etc make the search problem itself simpler. As the dog progresses with a good start, a powerful search, and a clear, strong alert each time, you can begin to challenge the dog by changing only one or at most two variables at a time. If you change two variables, change one of them only slightly at first. Be sure to set the dog up for success each time by incrementally changing the problem, and making sure the reward is given to the dog to keep the responses strong.